Customer Experience Matrix

What Comes After the Composable CDP?

2025-12-15T16:42:00.001-05:00

For a bunch of smart people, the technology community can be awfully slow learners. One mistake they keep repeating is believing that the latest technology is a “silver bullet” that solves all problems. This despite years of experience finding that (a) no technology solves all problems and (b) every new technology is eventually replaced by something else.

The Customer Data Platform industry is as prone to these errors as everyone else. The current silver bullet is composability, which is touted as the ultimate solution to problems of CDP cost and complexity. Just hearing the claim should lead anyone to realize it’s not true: composability solves some problems in some situations (i.e., unnecessary data movement at firms with sophisticated technical resources) but it won’t help companies that lack the necessary data infrastructure or staff resources. I suspect that even the most ardent composable enthusiasts would agree with this more nuanced view if you pinned them to a wall.

It's a step in the right direction to move past “composability is great” to discussing “when composability is the best solution.” But it’s also important to address the second flaw in the “silver bullet” fantasy, which is to assume that the best current technology is the best future technology. In other words, it’s worth asking, what comes next?

Remember, composability isn’t a solution: it’s an architecture. The fundamental argument for composable CDPs is that it’s better to have an architecture that uses the enterprise data warehouse as the primary store for customer data, rather than an architecture where customer data is assembled, stored and accessed in a separate CDP database. So the question about what’s next is really a question of what other architectures are possible.

I can think of two candidates.

Read customer data directly from the source systems, without assembling it in any persistent database (i.e., neither a data warehouse nor a CDP). This vision of real-time, on-demand customer profile assembly is what many people thought a CDP should be back in the early days of the industry. At that time, ten to fifteen years ago, the existing technology simply couldn’t make it work. Many source systems didn’t allow real-time access, internal networks couldn’t handle the volume, and the processing costs to assemble profiles in real-time were too high. The problem pre-dates CDPs, of course: the need to make data accessible by pulling it from source systems into a separate database is why data warehouses were first created in the 1980s. There has certainly been progress in recent years, although I can’t say that the necessary technology is fully available. Still, the scope of what can be done in real time continues to expand, which diminishes the share of data that must be assembled in a warehouse. This, in turn, diminishes the relevance of the warehouse-centric composable architecture.

Create customer digital twins. I’m highly skeptical that AI can accurately mimic customer behavior: this seems most likely to reinforce mediocrity by failing to predict unexpected behaviors (the most interesting kind) or reactions to novel situations (which includes the most important innovations). But I can still imagine creating a digital twin for each customer, updated in real time with data captured across all company systems. The twins would be self-contained objects (or maybe agents) that can be queried any time a company system wants to find the best action for a particular customer in a given situation. While their role might resemble the customer profiles assembled in today’s data warehouses, I’m sure the technology will be quite different. I can't predict the specifics of that technology but am sure that clever people somewhere are already working on it. My only point here is that, again, the architecture won’t look like the warehouse-based composable CDP.

There are surely other possibilities that haven’t occurred to me. In fact, I’d offer better than 50/50 odds that the next CDP silver bullet will be something I haven’t listed. Placing the right bet matters a lot to investors and developers but not so much to users, who can wait to see what becomes available.

What does matter to users is identifying their requirements. This is something that debates about architecture only distract them from doing. So my fundamental recommendation is that CDP users think less about chasing the silver bullet and more about building the golden record – that is, how to create the accurate, unified, accessible customer data that a CDP using any architecture is intended to provide.

Fitting Agents into the Sales and Marketing Mix

2025-12-12T16:06:00.001-05:00

Much has been written recently about how marketing and sales processes change when human buyers and sellers are replaced by buyer and seller agents: abbreviated, inevitably, as “A2A” marketing. It’s a fascinating topic but just one model that will coexist in the near future with human (or, more precisely, non-agentic) buyers interacting with agentic sellers, agentic buyers interacting with human sellers, and, lest we forget, humans interacting with humans. Any consultant will immediately recognize that this cries out for a 2x2 matrix, or perhaps a pair of 2x2 matrices if you want to distinguish business marketing from consumer marketing. For the moment, let’s stick with the single matrix model:

It’s worth making these admittedly-obvious distinctions because each situation raises separate issues, which are otherwise easily jumbled into a confusing heap. Let’s look at each situation in turn.

Human to Human (H2H)

Beyond the literal situation of one seller talking to one buyer, I’d argue this also includes humans interacting with traditional broadcast media, web search, and even non-agent websites. The common thread is that the human buyer does most of the work of asking questions and processing answers. The seller is largely reactive, although there are some situations where she makes choices such as selecting a personalized “next best action”, embedding dynamic content in a website, and setting up conventional search engine optimization. Those choices may be informed by predictive models or some other type of AI, but every step in the workflow is ultimately managed by humans, not agents.

I can’t point to specific data but am pretty sure that H2H interactions still account for the vast majority of today’s sales and marketing activity. This means that marketing and sales teams should still give significant amounts of attention to improving them, even though agentic interactions are vastly more fun to think about. If you absolutely must bring AI and agents into the picture, you can use them behind the scenes to speed up workflows, optimize performance, and analyze results.

Agentic Buyers to Human Sellers (A2H)

This is probably the situation that gets the most attention today. It includes true “buyer agents” (controlled directly by buyers) and “buyer-supporting” agents such as AI search engines and browsers. I call these “buyer-supporting” because they’re not controlled by the buyer, but instead by a company like OpenAI or Google which provides them to buyers at little or no cost.

The distinction matters because companies that offer “buyer-supporting” agents have their own agendas, which don’t necessarily align with the interests of actual buyers. In particular, these companies are increasingly interested in monetizing their products by serving ads within AI search and browser results. Some of these ads will be clearly labeled while others may be subtly embedded in the results themselves. These ads are an opportunity for marketers but may be problematic for users, who could be led to question the objectivity of the AI results.

Concern about biased AI search results could in turn lead to significant interest in true “buyer agents” that consumers pay for themselves. History suggests this will be an uphill battle: as we’ve seen with streaming video, large majorities of consumers typically chose free, ad-supported services over paid, ad-free subscriptions. Still, as streaming video has also shown, a significant fraction of consumers will pay for subscriptions in return for a better experience. This could be a large enough market to support a profitable business. Business buyers are even more likely to purchase agent subscriptions, since they don’t pay with their own money and can easily justify the expense based on better quality results. The precedent here is ad-supported versions of office productivity apps, which have never been broadly successful. There’s a chance that agents could be funded by charging advertisers for access to their owners, although such models have also failed in the past.

Advertising aside, most A2H discussions in martech and adtech circles focus on how sellers can adapt their systems to get the best results from buyer-side agents. This often involves advice on optimizing website design to accommodate search and browser agents, so a given brand receives the best possible treatment. Traditional SEO vendors are frantically expanding their products to meet this need and new AEO (AI Engine Optimization) specialists are also appearing. So far, the solutions are pretty basic: systems run sample queries to measure how often a given brand is mentioned in AI search results and vendors offer design tips to expose the kinds of data that AI agents are looking for. The next level is to look beyond measuring and influencing whether the brand is presented, to how it’s presented in terms of positioning and value. We’ll surely see more of that.

The thing to remember about “buyer-supporting” AI search and browser agents is they are generally driven by a big LLM model that draws from the same information for all users. True “buyer agents” would supplement the more-or-less static LLM models with custom research that visits seller websites to find answers to buyers’ specific questions. For example, one buyer might be interested in pricing details while another cares more about product quality. Beyond exposing all possible information, a seller might aim to present its product differently depending on what appear to be the buyer’s priorities. This is largely similar to today’s (non-agentic) website personalization. What’s more intriguing is the possibility that sellers can find a way to identify individual buyers’ agents over time, perhaps by requiring registration in exchange for detailed information. This would let the seller build a buyer profile and tailor responses to this profile. Piercing the buyer agents’ veil of anonymity would be hugely valuable.

There is a third situation: where the “H” in “A2H” is an actual human, not a non-agentic system. One current example is humans responding to agent-generated Requests for Proposals, which will likely be joined by other formats such as email inquiries or even telephone surveys. The growing volume of agent-generated requests is already a nightmare for business sellers faced with the cost of responding to them. The obvious solution is to let seller agents respond to the buyer agents, but it may be a while before most firms can deploy this capability. In the interim, sellers will be increasingly pressed to qualify buyers before deciding how to respond. Insofar as responding to qualification questions requires effort by the buyer, this imposes a cost on the buyer that should help to eliminate frivolous requests. At some point it might make sense for sellers to impose a literal cost – that is, to charge a fee – for responding to agent-generated sales queries. A less obvious concern is that buyers who rely on agent-generated research questions may fail to understand their true needs, removing a substantial portion of the value gained from a good purchasing project.

Human Buyers to Agentic Sellers (H2A)

Traditional websites may use AI-driven personalization but they are still non-agentic systems. In the future, we can expect true agentic interactions to become increasingly common. The best current example would be chat interfaces connected to an agentic back-end, enabling them to engage in true conversations with potential buyers. These have already evolved in some situations to full-scale agentic business development reps (who send those those super-annoying emails complementing your latest blog post and asking for an appointment) and sales reps (engaging in lengthy dialogs). Agentic customer support reps are even more common and, often, better than humans at many tasks. While the distinction between AI-based and agent-based interactions can be vague, it’s fair to say that agentic interactions will be significantly more responsive to individual situations. This, in turn, makes them more reliant on capturing real-time data, both for customer behaviors and surrounding context.

Letting autonomous agents interact directly with customers raises major concerns about governance, output quality, and risk. These are widely recognized, as are the challenges of integrating agent-based systems with existing infrastructure. That being the case, I won’t rehash them here, apart from noting that they currently present substantial barriers to adoption of H2A models.

Agentic Buyers to Agentic Sellers (A2A)

Agents selling to other agents is the obvious endpoint of agentic adoption. It’s appealing if only for the amusing prospect of agents merrily jabbering with each other without any human involvement. But apart from a few highly structured interactions, such as programmatic advertising, it’s still largely in the future. A2A can’t become more common until the industry first solves the separate challenges of agentic buyers and agentic sellers. It must then overcome the additional challenges of connecting the two. Once the plumbing issues are addressed, there will be another level of adoption as buyers and sellers work to turn the interactions to their advantage. How will price negotiations work when buyers want the lowest price possible and sellers want the highest price? How will sellers discover the actual needs of buyers so they can make the best recommendations – and is what’s best for the seller necessarily what’s best for the buyer? How will seller agents decide which information to offer and which to exclude? How will agents build trust with each other? And how will companies manage the computing costs of agent-to-agent interactions, which could be substantial if the interactions are extensive?

Plenty of smart people are surely working through these issues. We already see some technical foundations being laid in protocols such as MCP and Google’s A2A. But it’s probably too soon for most marketers to put much energy into worrying about A2A deployment. Mastering the intermediate steps of A2H and H2A should come first and will put them in a better position to deal with A2A when the time is right.

Summary

The impact of AI in general, and agentic AI in particular, is overwhelming. While this piece offers some ideas and makes some prediction, my real goal is much simpler: to suggest that distinguishing the different types of human and agent interactions is a way to split the topic into smaller, more tractable pieces. I hope that helps.

Let's Debate CDP Functions, Not Definitions

2025-11-14T17:52:00.005-05:00

What’s the definition of a CDP? It's a bad question because it diverts attention from what really matters: What capabilities do CDP users need? Still, buyers keep asking and sellers keep answering, typically in ways that promote their own interests. Looking for an unbiased perspective on the topic, I recently asked ChatGPT what answers it was finding. It came back with a reasonable cluster of responses and particularly interesting details on who was using each one (see below for the full response)*:

Unified customer database: 70–90% of analyst pages, trade articles and vendor docs
Marketing activation / audience building platform: 60–85% of vendor docs, blogs and many press releases
Real-time/streaming profile & interaction engine: 30–60% depending on whether the source is vendor marketing (more likely) or neutral analyst articles (less likely to require “real-time”).
Privacy, governance & identity management layer: 15–35%, increasingly present in analyst pieces and vendor positioning
Part of a larger ‘data cloud’ or enterprise data stack: 10–30%, especially in vendor/marketing copy from big cloud vendors

These are categories that ChatGPT identified without me defining them in advance. So it's particularly interesting that there’s no mention of composability, warehouse-centricity, no-copy, hybrid, embedded, integrated, stand-alone, or other architectural details that have dominated recent industry discussions. In a way, this represents a failure by the CDP Institute to propagate our view that a CDP must build a separate database of its own. But a less parochial response is to be pleased that the main distinctions reflect system functions, which is where the focus belongs.

Of course, the variety of definitions is still problematic. While it’s usually safe to assume that a system labeled as a CDP will provide a unified customer database, it’s less certain that it will also offer marketing activation and downright dicey as to whether it will offer real-time streaming profiles and interactions. This means the label provides little useful guidance: imagine a can of soup labelled “contains tomatoes and maybe chicken and could also have mushrooms, rice or shrimp”. The only way to know what's inside would be to open it -- which defeats the purpose of a label.

(And, yes, the problem is worse for CDP than other categories. When I ran the same prompt for the term "customer relationship management software," a single answer dominated: 71% defined CRM as “a software/system/platform to manage customer interactions and data.” The next highest share was just 29% for “integrated suite (sales, marketing, service automation)”. It’s true that the dominant answer is exceptionally broad, but at least most people understand this and won’t expect anything more specific.)

So, although industry understanding has not been entirely destroyed by architectural debates, there is still enough disagreement on the scope of a CDP to limit the term’s utility. (If the CRM example is typical, it may be a natural progression for popular categories to expand their meaning over time. That would be an interesting hypothesis to explore if anyone out there is looking for a thesis topic.)

The industry could fight to restore a more specific CDP definition, but that’s probably a losing battle. It’s more likely productive to shift the discussion away from defining the term "CDP" to defining the functions required to manage customer data.

Yes, I’m proposing that the solution to our problem is a checklist. Don’t roll your eyes: whole books have been written on the topic. (Ok, maybe just one book.) But in an industry that has long been driven more by theory than practical requirements, anything that gets buyers to focus on what they actually need is a win.

Getting the industry to agree on a shared requirements checklist wouldn’t be easy, since every participant would want to add or remove items depending on whether their products supports them. Indeed, the very notion of a comprehensive requirements list favors broad, integrated products over narrow point solutions. But I’d still invest a few embers of hope in a project to forge a complete customer data framework. The potential benefits, for users and vendors alike, are well worth the effort.

*Here's the ChatGPT dialog:

Prompt: what are some common definitions for the term 'customer data platform'? Indicate how frequently the different definitions appear in sources such as trade publications, press releases, blogs, and social media posts.

Response: I sampled authoritative analyst sites, vendor docs, trade publications and typical vendor/blog explainers to pull together the common ways people define “Customer Data Platform (CDP)”, then estimated how often each definition shows up across those source types (trade press/analyst pieces, vendor press releases/docs, blogs, social posts). Below are the common definitions (short label + one-line meaning), an evidence-backed frequency estimate, and the strongest citations.

________________________________________

Common definitions of “Customer Data Platform” (with frequency estimates)

1. “Unified customer database” (the CDP core definition)

What it means: Packaged software that builds a persistent, unified customer database (stitches identity across sources) and makes that single customer view accessible to other systems.

How often it appears: Very common — ~70–90% of analyst pages, trade articles and vendor docs emphasize this as the defining feature.

Why / sources: This is the CDP Institute’s explicit definition and is echoed by major analyst glossaries. (CDP Institute)

2. “Marketing activation / audience building platform”

What it means: A system that not only unifies data but builds audiences/segments and activates them (syncs to email, ad networks, personalization engines) for marketing campaigns.

How often it appears: Very common — ~60–85% of vendor docs, blogs and many press releases emphasize activation as a key purpose.

Why / sources: Vendors and marketing-focused guides (Segment, Hightouch, Salesforce) present CDPs in these terms. (Segment)

3. “Real-time/streaming profile & interaction engine”

What it means: Emphasis on real-time ingestion, updating of profiles and immediate activation for personalization / real-time interactions.

How often it appears: Common — ~30–60% depending on whether the source is vendor marketing (more likely) or neutral analyst articles (less likely to require “real-time”). Many vendors promote “real-time CDP” capabilities.

Why / sources: Adobe, Salesforce and many vendor pages highlight real-time processing as a differentiator. (Adobe for Business)

4. “Privacy, governance & identity management layer”

What it means: CDP framed as a way to centralize consent, governance, and identity resolution for compliance and safer data sharing.

How often it appears: Occasional — ~15–35%, increasingly present in analyst pieces and vendor positioning as privacy/regulation matters grow.

Why / sources: CDP Institute discussions and recent vendor guides mention governance as important but not always the central definition. (CDP Institute)

5. “Part of a larger ‘data cloud’ or enterprise data stack” (broader/platform definition)

What it means: Some sources treat CDP as overlapping with or evolving into broader data-cloud/data-lake capabilities (enterprise data cloud that supports analytics, AI, and cross-org use, not only marketing).

How often it appears: Rising / context-dependent — ~10–30%, especially in vendor/marketing copy from big cloud vendors that bundle CDP features with broader data clouds.

Why / sources: Salesforce/Adobe messaging and some modern vendor blogs frame CDP within a larger data-cloud narrative. (Salesforce)

6. “Narrow ‘marketing-only’ vs broad platform debate”

What it means: Some authors contrast CDPs that are strictly marketer-focused (audience building + activation) with systems that are more IT/analytics-focused (which some call Customer Data Infrastructure rather than CDP).

How often it appears: Common in commentary/blogs — ~25–50%, especially in independent analyst/blog posts arguing about scope.

Why / sources: Independent guides and blog explainers highlight this debate. (Salesforce Ben)

________________________________________

Quick guide to where each definition shows up most often

• Trade publications & analyst glossaries (Gartner, CDP Institute, Forrester pieces): emphasize unified database, identity resolution and standard criteria (CDP Institute/Gartner). (CDP Institute)

• Vendor press releases / product pages (Salesforce, Adobe, Segment, Hightouch): emphasize activation, real-time processing, and enterprise/marketing use cases — vendors often broaden the definition to include features they sell. (Salesforce)

• Blogs / explainers (independent and vendor blogs): often present a practical, functional definition (“collect, unify, activate”) and discuss implementation tradeoffs; they also host the “marketing vs data-platform” debate. (hightouch.com)

• Social media (Twitter/LinkedIn threads, product announcements): shorter, use-case driven framings (e.g., “CDP = audience building + sync”), vendor threads hype product differentiators (realtime, privacy). Frequency estimates are noisier here but themes mirror vendor/blog messaging.

________________________________________

Methods & confidence

• Method: I sampled canonical analyst definitions (CDP Institute, Gartner), major vendor pages (Salesforce, Adobe, Segment, Hightouch), and representative blog/explainer pages. I then categorized recurring definition themes and estimated how often each theme appears across source types.

• Confidence: Moderate–high for the major categories (unified database; activation for marketing), moderate for newer/marketing-spin categories (real-time emphasis; data-cloud framing). The percentage figures above are approximate — they’re meant as quick directional estimates, not exact counts — because the corpus (web) is large and vendor language shifts fast.

________________________________________

Short takeaway (one sentence)

Most authoritative definitions center on a CDP as packaged software that creates a persistent, unified customer database that marketers and systems can use — with many vendors and blogs adding that a CDP also builds audiences and activates them (often in real time), and some vendors now frame CDPs as part of larger “data cloud” offerings. (CDP Institute)

Why Agentic Campaigns Are Not The Future of Marketing

2025-11-02T16:03:00.002-05:00

My last post, presenting ChatGPT’s description of AI-based marketing, generated thought-provoking comments on LinkedIn. Key insights included:

Organization issues as roadblocks to successful deployment (Chris Adelman and Lee Hammond)
Buyer agents as increasing the importance of trusted information sources (Jon Miller, Dan Cote)
Some vendors are already delivering end-to-end agentic campaigns (Matthew Niederberger, Duarte Garrido)

Thanks to everyone who contributed.

What I didn’t see was push-back against the vision that ChatGPT presented. Maybe people were being polite or maybe they fear that ChatGPT could hold a grudge. But my own initial take was triggered by words including “narrative” and “journey”. These led me to think that ChatGPT was still describing today's standard of multi-step marketing campaigns applied to customer segments. On closer examination, the story is a bit more complicated: the various components included in ChatGPT’s list are not necessarily a consistent whole, but rather a collection of predictions that ChatGPT has gathered by scraping of industry discussions. Oddly enough, ChatGPT's summary provides a clearer statement of an alternative approach:

AI-native marketing is:

Continuous (not campaign-based)
Conversational (not broadcast)
Collaborative (AI agents on both sides)
Generative (creating narratives, products, and experiences dynamically)
Ethically-audited and explainable (trust is as important as persuasion)

Of course, that ChatGPT says something is no guarantee that it’s correct. So, although I intuitively agree with the prediction, I still need to convince myself (and, hopefully, others) that it’s right. One test is to compare the prediction with cutting-edge industry developments. Here’s a list of trends I’ve been tracking:

Agentic search: AI-generated summaries replace traditional, link-based search results
Ads in chat replies: Chat responses include paid advertising, identified with varying degrees of transparency
Social commerce (a.k.a."live shopping"): Make purchases of items featured in social media discussions without leaving the social media platform
Commerce via agentic browsers: Advertising and shopping links presented by browsers based on what they see the user doing (a privacy nightmare but we'll ignore that for now)
Interactive ads: Advertisements that link to a conversational interface rather than conventional landing pages or websites
Buyer agents: AI agents that do product research and, perhaps, make purchases on behalf of users without users visiting seller websites
Conversational websites: websites that present chat-style interactions or hyper-personalized contents in dialog with the user

All but the last of these occur at the discovery stage of the buyer cycle. That's encouraging because marketers have been very concerned about they reach new customers when traditional search no longer drives traffic to their websites and consumers increasingly ignore traditional email and display messages. But what's really important is that these items all change the role of marketing from creating campaigns that broadcast messages to consumers, to executing continuous conversations with consumers. This is indeed what ChatGPT described. It’s not something that is accomplished by using AI generate conventional broadcast campaigns.

In fact, I’d argue that the current broadcast campaign model developed precisely because the then-existing technology couldn’t handle a conversational approach. Branching campaigns with personalized content nodes were the best that could be done when marketers were limited to deterministic, manually-crafted workflows. If you recall the once-popular cliché of marketing automation as a way to replicate the personal relationship of customers to their “corner grocer,” you quickly realize the real-life grocer had conversations, not campaigns.

Think a bit deeper and you see that the reason the grocer was able to have conversations was data. He knew the history of each customer and was able to discern their current situation by seeing and talking to them. Multi-step marketing funnels exist because marketers don’t have that kind of information available in real time. Instead, they have to make a guess at where each customer might be in the buying cycle. The funnel is a framework for making those guesses. Provide real information and it’s no longer needed.

I think this is important. Marketers, vendors and analysts have focused largely on how AI can reduce the cost and time of generating marketing campaigns. This allows companies to create near-infinite amounts of personalized content. But unless that content is driven by near-complete, real-time data, there’s a severe limit to how relevant it can be. Data, not content creation, is what ultimately limits the effectiveness of AI marketing.

This may strike you as bad news. Data has always been a problem for marketers, and real-time data is a particular challenge. Agentic campaign creation won’t solve the data problem, although perhaps it will highlight data's role as a bottleneck. The good news is that other AI and agent applications can in fact improve data access, something that data management vendors are already pursuing.

It’s true that better data management isn’t as exciting, or directly tied to revenue, as better customer engagement tools. This is probably why most vendors focus on adding engagement features. But as autonomous engagement become increasingly common, the need for better data will become more obvious. Vendors who use AI to deliver better data (and convince the market that they are leaders in doing it) will have an increasing advantage in growing their business.

I’ll give the final word to Abhi Yadav of iCustomer, one of those very smart people who see further, sooner than most of us. He recently shared a substack post that captures my opinion quite eloquently:

We used to drag customers to our properties (site, app, landing pages). Now we take capabilities to where customers already live: agent/app ecosystems, workflow-native GPTs, and search/ad surfaces that embed decisions at the edge. ... If your data stays scattered and your channels siloed while this shift accelerates, AI won’t transform your business it’ll just perform visual eye candy. AI should drive a paradigm shift in your business, not a party trick.

What would an AI-native marketing process look like? ChatGPT's answer is better than I expected

2025-10-30T14:30:00.005-04:00

If you look at recent vendor announcements, it seems that the cutting edge in applying AI to marketing is 'end to end' campaign generation, where a collection of agents work together to build and execute a complete marketing campaign with no human effort (although, for practical purposes, still under human supervision). As I've written before, this strikes me as an advanced type of 'substitution', where the new tech is used to execute the old processes better and faster. That's fine and valuable but it's only the first step in taking advantage of the new technology. The final step is to develop new, "transformed" processes that are optimzed to make full use of the new tech's capability.

I've recently been asking myself and others what they think that "transformed" process would look like for marketing. I'm still collecting my thoughts but, as part of the process, asked ChatGPT for its opinion. The answer was better than I expected. I don't necessarily agree with all of it but thought it was interesting enough to present to the larger world for comment. So, please, let me know what you think.

Prompt: if marketing techniques were designed from scratch to take full advantage of AI, what would they look like?

That’s an excellent and forward-thinking question — and the answer’s pretty transformative.

If we were to design marketing from the ground up for AI (instead of retrofitting AI into traditional marketing), we’d end up with a system that looks less like “campaigns” and more like continuously learning ecosystems that respond in real time to individuals and social dynamics.

Here’s a breakdown of what that might look like:

1. Autonomous, Adaptive Market Intelligence

Instead of periodic market research and segmentation,
AI-first marketing would run on constant sensing loops:

AI agents continuously scan social media, purchase data, and search patterns.
They detect cultural shifts, emerging memes, and micro-trends as they form.
Personas aren’t static — AI builds and dissolves them dynamically based on live data.

Analogy: You don’t define “target audiences.” You discover living patterns of attention in real time.

2. Hyper-Personalized Narrative Engines

Instead of a single ad creative for millions of viewers,
AI-first marketing generates unique narrative journeys per person:

Language models craft individualized story arcs and product framing that match someone’s identity, tone preferences, and current emotional state.
Multimodal models (text, image, video, audio) assemble personalized content on the fly — like an AI “showrunner” creating micro-stories for each user.
Brand consistency is enforced by a central “brand DNA model” ensuring all outputs feel coherent and on-brand, even when autonomously generated.

3. Predictive and Generative Commerce

Instead of advertising then selling,
AI-first marketing anticipates needs and builds products or bundles before people articulate them:

Predictive demand modeling uses behavioral and environmental data to suggest what products should exist.
AI may prototype and A/B-test new offerings via virtual storefronts, gauging interest before production.
Marketing and product design merge — you market possibilities before they exist, and only produce what the market proves it wants.

4. Multi-Agent Relationship Systems

Instead of one-way brand broadcasts,
AI-first marketing uses relationship agents that represent both consumers and brands:

Consumers might have personal “shopping AIs” negotiating with brand AIs.
Trust, transparency, and reputation become measurable assets traded between AIs.
The marketing goal shifts from persuasion to mutual optimization — aligning a person’s needs and a company’s offerings through agent collaboration.

5. Real-Time Ethics and Trust Architecture

AI-driven marketing would require built-in ethical governors:

Explainability modules ensure users understand why they’re being shown something.
Consent becomes dynamic and contextual (you can allow certain data uses for limited time).
Bias-detection systems continuously audit generative campaigns for fairness and cultural sensitivity.

6. Emotionally Intelligent Interaction

Instead of static messaging,
AI-first marketing converses, empathizes, and co-creates:

Emotional state detection (via tone, text, or visual cues) guides responses.
The AI adjusts its persuasive approach — sometimes offering information, sometimes humor, sometimes retreat — like a human skilled in timing and empathy.
Brands become more like adaptive characters than faceless institutions.

7. Closed Feedback Loop: Learn → Generate → Test → Learn

Every interaction feeds a continuous reinforcement loop:

Each ad, email, or product suggestion is an experiment.
The system measures micro-responses (hover time, tone change, conversion intent).
Those signals retrain the creative and strategy models daily.

In summary

AI-native marketing is:

Continuous (not campaign-based)
Conversational (not broadcast)
Collaborative (AI agents on both sides)
Generative (creating narratives, products, and experiences dynamically)
Ethically-audited and explainable (trust is as important as persuasion)

How Will GenAI Change the Consulting Industry?

2025-10-26T14:06:00.009-04:00

Reports on the threat that generative AI poses to knowledge workers usually focus what might be called “structured creativity”: tasks like copywriting or computer programming where outputs are clearly specified through things like creative briefs or system requirements. One group that generally doesn’t consider itself at high risk are the analysts and consultants who write the reports themselves. After all, their work requires insights and judgements that are beyond the current capabilities of genAI, which ultimately does nothing more than play back the training data it's fed.

Case in point: This widely reported BCG study found that genAI tools substantially improved performance of consultants at product planning, with the greatest improvements among below-average performers. A second finding of the same study was less widely noted: AI depressed performance at generating insights from interviews and raw data. This dovetails nicely with the idea that AI can help with structured tasks such as campaign management but isn’t (yet) capable of developing new insights. One reason that AI struggles with insights is its well-known tendency to hallucinate, which would translate into false conclusions about the meaning of input data. (See, for example this BBC study that found 47% of AI chatbot answers contained significant errors about data sources or accuracy.)*

So it seems that the survey analyzers are indeed safe, at least for the time being. But while insights are an important part of the value provided by analysts and consultants, they’re just a small fraction of their work. In my experience, the main reasons that companies hire consultants are to access their skills at organizing a project and their knowledge of vendors to consider hiring.** Those are things that genAI does extremely well: ChatGPT will gladly provide a plausible project plan for nearly any task and will just as merrily answer your questions about which vendors to consider, providing detailed comparisons on relevant features. It's true that those comparisons will often be inaccurate (see the BBC study or do your own test in a field you know well). But, let’s face it, consultants also make mistakes about product capabilities, which change much faster than even the most diligent industry expert can keep up. So, whether a vendor list comes from genAI or a human consultant, it still has to be validated with current information before you base any decisions on it.

Even more to the point: just as it’s always been hard for clients to tell truly knowledgeable consultants from pretenders, it’s now equally hard to tell accurate genAI outputs from hallucinations. This applies to frameworks, vendor lists, and even those insights where humans are still clearly superior. GenAI is famously capable of providing reasonable-sounding arguments based on false premises, citing false evidence, and leading to wrong conclusions. A true optimist would argue those can lead to fruitful discussions even when they start with an error. A more realistic expectation is that some large fraction of people will accept them as delivered.

This is especially true when it comes to comments on industry trends and futures, where there is no clearly right answer. A large part of the industry conversation today, especially in media like LinkedIn, is exactly that sort of discussion. AI-generated comments are increasingly common and, truth be told, can be hard to tell from the human-generated ones. Sure, that's a tough situation for industry members trying to get good information. But save some sympathy for the poor human consultants who are trying to be heard – and recognized as legitimate experts – amid all the noise.

Indeed, the flood of content is arguably the biggest change in marketing in recent years. Not so long ago, publishing good content on a regular basis was an effective, low-budget path to success. The mechanism was simple: good content was picked up via social media sharing and, if it became popular enough, registered in search engine results. Today, not only is there too much content for anyone who’s not already famous to gain a significant share of attention, but AI search summaries will scoop up good information without leading readers back to the source. The alternative path of publishing on high-traffic news sites is equally blocked since AI search engines stifle traffic to those sites as well. Of course, a handful of influencers who are skilled at attracting attention can still become Internet-famous. But what gains attention isn’t necessarily what’s of true value.***

I’ll skip the usual moaning about bad information driving out the good, since the consequences are all around us. Instead, I’ll posit that the situation is untenable because people simply will not be willing to consume such huge volumes of content indefinitely. There is already a backlash among people who want to reduce the time they spend trying to keep up. (In fact, time spent on social media peaked in 2022.) The obvious solution is to let AI view media for you and summarize what’s worth seeing. Some people may find that acceptable but I’d like to think that most will recognize that AI can’t be trusted to make the right choices – especially if the AI is run by corporations with interests of their own. The alternative would seem to be relying on human curators – remember “editors?” -- to assess the available materials and select those worth consuming. Of course, deciding which curators to trust will itself require consumers to do some work. This path also requires an alternative to the failing economics of traditional, ad-supported media, which will probably be consumer-paid subscriptions that compensate both the content creators and the curators. A middle ground might be AI curators that are funded by consumers rather than self-interested corporate providers.

I won’t pretend to know approach will dominate. Perhaps different groups will adopt different approaches, with access to accurate information becoming a luxury good like ad-free TV. For now, people can already be selective in their content consumption by avoiding the undifferentiated (or, more precisely, platform-selected) streams in news and social media feeds. The dilemma is sharper for experts relying on content creation to build their business: do they try to keep up with content bots that can easily outproduce them, recognizing that even consistently brilliant, unique, and insightful content won’t necessarily be noticed? Or do they find other ways to reach prospective clients, perhaps through live events or less mediated platforms like reddit and medium?

Mind you, this is occurring in a world where AI can legitimately replace some of the expertise consultants are used to selling, such as methodologies, technical skills, and vendor knowledge. This leads to smaller engagements where clients are buying only the AI oversight and creative insights that humans alone can provide. So does the total market for consulting services shrink at exactly the same time that selling those services becomes harder? Or, more optimistically, does lower cost enable consultants to expand the market by selling to firms that couldn’t previously afford their services? (See, for example, this new WPP offering of AI-enabled “self-service” ad campaign development.) While AI-powered services don’t themselves create more human jobs, they do presumably offer an opportunity to sell insight-based projects that do require paying humans.

I'll say it again: the future is unknown. But I do believe that analysts who are calmly assessing the threat that AI poses to other people’s jobs might give more urgent thought to the threat AI poses to their own.

* I’ll ignore the distinction between AI augmenting workers and AI replacing them: If you make people twice as productive, then half the workers are no longer needed. It’s true that there’s some added work to check AI results, but, despite complaints about ‘workslop’, we can probably assume that there’s a substantial net labor savings. Certainly most AI users say they’re satisfied with the benefits despite the extra work.

** Caveat: my experience is biased since most of my consulting was around vendor selection.

*** My entry in the Understatement of the Year competition.

Taking Advantage of AI Search

2025-08-23T12:40:00.002-04:00

The past few weeks has seen a flurry – more of a blizzard, really – of announcements from Search Engine Optimization companies that have added tools for AI search ranking. This is a natural outcome of the meteoric growth in AI search: referrals from ChatGPT grew 20% from June to July this year while ChatGPT agent visits doubled in the same period.

Before you get too excited, bear in mind that LLM vendors’ share of global search is still a minuscule: 0.13% so far in 2025 (for ChatGPT and Perplexity combined). Given that Google still has nearly 90% of all search traffic, it’s much more important that more than half of Google searches now return an AI overview.

Probably the most important statistic is that organic click-through rates drop by 20% to 40% when AI Overviews are displayed. One important caveat is that AI overviews show up primarily on “informational” queries, as opposed to commercial, transactional, or navigational queries. The caveat to the caveat is that informational queries are by far the most common: 88.1 % of the total, compared with 8.7% commercial, 1.4% transactional and 1.8% navigational. So, if your web strategy has been based on providing information to attract visitors, you have some adjusting to do.

Or do you? Most AI optimization tools replicate SEO by focusing on how often a brand appears in AI search results. This Evertune press release provides a detailed explanation of how their product works, which is probably typical: the system runs a set of common prompts on a topic and tracks how often the target brand is mentioned. That’s very similar to traditional SEO, which reports how the target brand ranks in search results. In both cases, the goal is to be seen more often.

It's true that tactics to reach those goals will differ. Traditional SEO is based on links and keywords, while AI search engines care more about detailed information and reviews. More broadly, content that attracts humans will often be insightful or entertaining, while bots respond best to clearly presented data. So some changes are needed if you want to switch your goal from SEO to AI search ranking.

The deeper question is whether appearing in AI search results is the right goal. If you’re a brand whose aim is building awareness and credibility, the answer is probably yes – so long as the AI results present you in a positive light. It will be even more important as search bots morph into shopper bots, which recommend (and will eventually execute) purchases on behalf of human buyers. Shopper bots will no doubt engage with brand-owned seller bots, a category (or e-commerce system feature) that I can't recall seeing but is surely gestating as you read this.

The answer is different if you’re a publisher. There’s little value in having your materials cited in an AI response that doesn’t generate a click-through to your site. In fact, showing your materials gives users less reason to visit. Adding injury to insult, responding to the AI search engine’s crawler actually adds to your website operation costs. Put simply, publishers need traffic. Visitors provide inventory for ad impressions and names to capture for future promotions. The drop in organic traffic forces publishers to seek other traffic sources, including both paid search and social forums such as reddit, Instagram, and YouTube.

Paid search itself will evolve along with AI search. Ads already appear alongside Google and Microsoft AI search results. Looking ahead, Google Research recently published a paper describing how advertisers could bid on the text of the responses themselves. There are also other ways to reward AI search engines for driving traffic to advertisers, such as affiliate relationships. Consumers won’t be pleased when paid and unpaid recommendations mingle indistinguishably in search results, but it’s not clear they’ll have much choice in the matter. The problem facing publishers is that many won’t be able to afford the cost of paid advertising or will need to reengineer their business to generate revenues to cover the expense.

Social traffic is a more appealing option for publishers since it’s “free” to the degree they can generate it. That should be possible: after all, publishers' core competency is to create appealing content. One implication is a growing overlap between traditional publishers and influencers, a distinction that has already become quite blurry. It’s worth noting that we’ve recently seen substantial growth in software and services to support influencer marketing, including tools to find influencers, build audiences, and streamline content creation, as well as ad networks and media buying tools to simplify ad placement on influencer sites. Most have been aimed at individual influencers, but enterprise-grade products will probably appear with publishers as the target clients. Social marketing will increasingly look like SEO, except aimed at optimizing content for social algorithms rather than search algorithms. Since SEO companies and social media management vendors both have a legitimate claim to this turf, it will be interesting to see who ends up owning it.

In short, we’re just beginning to see the changes resulting from AI search. It has already disrupted the business model of many web publishers, who no longer receive the organic search traffic that powered their audience creation. Traditional SEO will be supplemented by AI SEO, especially for brands who can benefit from appearing in AI search results. But while traditional search generated large volumes of free traffic, AI search results will likely be dominated by paid placements. Publishers, who gain less than brands from AI search ranking, will need to find alternative sources of traffic and will increasingly resemble influencer business models. Above all, we’ll see continued development of supporting technologies (“infrastructure” in terms of the STIB model) as entrepreneurs identify new opportunities to help companies operate in the new environment. Ultimately, again according to STIB, we’ll start to see new business models emerge that are best suited to the dynamics of an AI search-based economy.

CDP Relay42 Bought by Supermetrics: The Embedded CDP Trend Continues

2025-06-18T17:04:00.004-04:00

After a flurry of CDP acquisitions in January, things had been quiet on that front – until this morning, when we learned that Relay42 is being purchased by Supermetrics. The deal reverses the pattern of earlier CDP acquisitions, which had been been by customer-facing companies in content management (Contentstack/Lytics), customer service agents (Uniphore/ActionIQ), and ecommerce (Rokt/mParticle). Earlier CDP acquistions had also largely involved customer-facing businesses such as retail marketing (Bloomreach/Exponea), email (Sitecore/Boxever) and messaging (Twilio/Segment). But bear in mind that Relay42 is what we call a campaign CDP, meaning it offers journey orchestration, personalization, and outbound campaigns, as well as advanced analytics and AI agents. So the underlying logic is the same: vendors want to provide both data assembly and activation. While other acquisitions added a CDP’s data assembly capabilities to an existing activation system, Relay 42 adds activation to Supermetrics data assembly. Either way, you end up with a combination of both.

The trend towards embedded CDPs is now so clearly visible that it’s barely worth a mention. A more interesting question might be why Relay42 chose to sell itself at this moment: after all, as a bona fide CDP, it already had data assembly features similar to what Supermetrics provides. I suspect the answer here is brutally simple: money. Despite having an excellent product and marque accounts, Relay42 has struggled in recent years, dropping (per LinkedIn) from 69 employees in June 2023 to 50 as of June 2025. Over that same period, Supermetrics has grown a relatively modest 12% but, more important, its total headcount is now 407. At eight times the size of Relay42, Supermetrics provides the financial resources and customer base to power future growth. Without additional resources, Relay42 would have been hard-pressed to continue to develop its product, let alone make the sales and marketing investments needed to expand.

I had intended to write a longer post this week expanding on the trend towards embedded CDPs. I even had a nifty graphic prepared to illustrate the changes. (Actually, I prepared it for a workshop I’m giving in July at the Vibe Marketing Tech Fest in Manchester, UK, if you'd like to attend.) But plenty has already been written on this topic and I’m not sure a more detailed exploration would add much value. The only point I haven’t already made is that the trend towards integrated customer data/activation systems runs counter to the trend towards assembling systems from separate components. At the risk of being boring, I’ll mildly suggest that both will find some takers. What the Relay42 deal does illustrate is that small, independent CDPs will find it very hard in the future to compete outside of niche markets.

Mapping AI-Based Martech Using the STIB Innovation Diffusion Model

2025-05-24T20:38:00.187-04:00

TL;DR: This post fleshes out the STIB model of innovation diffusion, tests it by classifying recent customer management product announcements into STIB categories, and draws some insights about the model and customer management trends. As a bonus, it proposes a future of "hyper-personalized, financially-optimized treatment of each customer during each interaction across all channels."

The Substitution-Transformation-Infrastructure-Business Model (STIB) model of innovation diffusion that I presented in my last blog post is intriguing, but does it reflect reality? One measure is whether it provides a useful classification of industry products. To test this, I looked at the past ten weeks of product news in the CDP Institute Daily Newsletter, which came to about 50 items. Results were quite interesting.

But before we get to that, let’s flesh out the model a bit. It describes two core elements:

Innovation: the development whose impact is being measured. This might be electric motors or internal combustion engines. (More formally, innovation is a change in technology, where technology is defined as the tools, methods and knowledge used to perform a task. "Technology" in this sense isn't always "technical": a new marketing methods and financial tools can be important innovations. But I digress.)

Application: the process or product changed by the innovation being studied. This might be manufacturing in factories (a process) or transportation vehicles (a product). The application is assembled from components, which might be physical objects in a product or steps in a process workflow. The same innovation might affect many different applications but it's often more useful to study just one.

The model describes how an application changes in response to an innovation. This change comes in three stages:

Substitution: the innovation is deployed as an exact replacement for one component in the applications, without changing anything else. Electric motors were hooked to turn factory shaftwork instead of water wheels; internal combustion engines replaced horses for powering carriages.

Intermediate States: this is a period of experimentation. Multiple components are changed as the industry explores how to best take advantage of the innovation. The innovation itself continues to mature during this period, opening up additional possibilities.

Transformation: the industry settles on an optimal design incorporating the innovation. This "final" design is fundamentally stable although incremental improvements continue. Factory tools incorporate built-in electric motors; motor vehicles locate the engine and drive train most efficiently.

The model also tracks changes over time to technologies beyond the application itself:

Infrastructure: this includes products, suppliers, and processes that support the transformed application or are changed as a result. Electric motors require generating stations, power grids, and new machine tools, while they result in more efficient factory designs. Parts and materials manufacturers support motor-driven vehicles while auto ownership spawns new roads, filling stations, employment opportunities, legal frameworks, and more.

Business model: this describes business methods such as revenue sources, pricing models, marketing, distribution, customer service, funding, and ownership models. Larger factories lead to scale economies that favor large, national brands. Complex auto technology favors large organizations with deep resources for research, mass production, national advertising, and dealer networks.

The graphic illustrates relationships among these elements.*

In the initial state, the application, infrastructure and business model are all in their baseline (i.e., pre-innovation) configuration.

This is followed by substitution, where the innovation is applied to make a slight change in the application while infrastructure and business model are unaffected.

In the third stage, intermediate states, developers try variety of changes to the application and the infrastructure starts to mature in response to growing demands. Some will be a dead end (direct current electrical systems, steam powered cars). Others may hit on the transformed design before infrastructure is available to support them. Vendors may also fail if they don't change quickly enough to keep up with competition.

In the final stage, developers converge on a transformed configuration and application change comes more slowly. Some companies drop out or merge as the market consolidates. The infrastructure continues to evolve and new business models start to emerge.

The Future of Customer Management

Applying the STIB model requires specifying the application being changed and the innovation that is changing it. This might be tricky in some situations, but it’s pretty simple in the current context: the application is customer management while the innovation is artificial intelligence.

What’s not simple is predicting the form of the transformed application. Hindsight about factory equipment or auto design is easy, but we don’t know how customers will be managed in the future. This prediction isn't strictly necessary: we can simply observe changes as they occur without assessing their relation to the final state. But that assessment helps to organize the analysis and gives buyers and developers decide where to invest their resources.

Predicting the transformed state is a thought experiment: identify the constraints imposed by the current technology and imagine what the ideal design would look like if those constraints were removed (and replaced by whatever constraints the innovation imposes.)

As I see it, the dominant feature of customer management today is the number of separate steps that are performed by different people. There’s a creative team with a multi-step process to develop content, a targeting team with a multi-step process to select audience segments, a media team with a process to buy advertising, and multiple operations teams with multi-step processes to deliver messages via websites, email, social media, connected TV, digital video, games, out-of-home, podcasts, and elsewhere. The reason we have so many teams with so many steps is the limited capability of individual humans: each can be expert in only a narrow area, so many must work together to deliver a complete result. And, because they’re human, each person can only deliver a relatively small number of outputs over time. Thus, each output must apply to many people to deliver messages to everyone.

AI removes this limit. At least in theory, a single AI agent can expertly execute all steps in the messaging process, combining content creation, audience selection, media buying, and delivery. In practice, a single super-agent is less likely than a master agent that calls on specialized subordinate agents. But so long as those agents complete their work almost instantaneously, the process will still function as a single step. Ideally, these agents would view all available data and collaborate with each other: content development would be informed by audience characteristics; media buying would take into consideration other channel opportunities; and so on. All these decisions would be coordinated to produce the highest total value: for a given advertising impression, the system(s) might simulate the results of several different creative treatments for several different offers for several different products, and then select the best creative/offer/product combination – or conclude that even the highest possible value for that impression is too low to justify the investment. This requires a sophisticated value prediction algorithm which would consider long-term as well as immediate results.

In addition to speed and coordination, the process would be completed at near-zero incremental cost.

In short, I see the transformed state as hyper-personalized, financially-optimized treatment of each customer during each interaction across all channels.

This is a radical change. (That's why they call it a transformation.) There are no more campaigns (defined as a predefined sequence of messages), no more audiences (defined as a group of customers who receive the same messages), no more content (defined as a fixed combination of text, images, and offers), and no more media buys (defined as group of impressions purchased together). What we have instead is a master agent orchestrating customer interactions across all channels to optimize use of budgets, data, customer attention, products, staff, and computing power. That agent would call internal and external data to guide its actions and would deliver messages across both internal (owned) and external (paid) media.

In the transformed world, today's creative, analytics, media, and operations departments have largely vanished, apart from a few human(?) experts left behind to monitor the AI. On the other hand, there might be more need for humans to do strategy and product development.

This vision also points to new infrastructure and business models:

Supporting infrastructure would include vendors building the base AI systems, which are too complicated for most companies to build for themselves (although non-technical users might use no/low-code tools to tune them); data tools to prepare, integrate, and expose data from all sources in real time; and analytical tools to measure interaction value.

Surrounding infrastructure would include new media that provide real-time access to their customers; low-friction integration methods to connect marketers with these systems; new billing and analytical methods; and social/legal frameworks to govern customer data collection, sharing and privacy.

New business models would be needed for companies and the systems, data, and media providers that support and surround them. With each interaction managed separately, providers might become on-demand services that charge for value provided in each case rather than billing based on labor, system use, or impressions.

I’m not sure this will happen. At most, I’d bet one of my later-born children on it. I’m laying it here out because the STIB model works best if you measure against a transformed state (and because it’s fun to think about.)

Mapping against the STIB Framework

That said, let’s try mapping recent product news against the STIB framework.

Application

The application we’re analyzing is customer management, which is roughly what we cover in the CDP Institute daily newsletter. This gives us a reasonable collection of announcements to work with, although it’s just a small sample based on items that happened to appear during a relatively brief period.

If the STIB model is correct, we should find clusters of products on a spectrum from business as usual, to substitution of AI within current processes, to complete transformation. Because transformation doesn’t happen all at once, we would expect to find several intermediate stages. Indeed, that’s the case.

Current State: The first cluster would hold products that execute the existing process without any change. AI-powered co-pilots might fit here. But co-pilots are so common that we don’t bother to write about them in the newsletter. Nor do we usually cover products that aren’t doing anything new. So this cluster is empty apart from one item about agentic helpers.

CommerceIQ Offers Agentic Teammates to Ecommerce Marketers

Substitution: The second cluster is simple substitution: products that replace a discrete task with an AI-generated equivalent. We see plenty of these, often offered as collections of (separate) AI tools for multiple tasks. We also see toolkits for users to build their own. So long as each AI tool executes one task separately from the others, this is still substitution. Recent examples are:

Intermediate Products: Now we move into products that change the underlying process but don’t reach the fully transformed state. The news items seem to fall into three clusters. It’s important to note that we didn’t define these in advance: they have emerged from the data itself.

Content Creation: This is a popular task to automate. It’s a single task but more than substitution because most vendors connect their content generator to response data and use this to automatically optimize content over time.

Goal-Driven Workflow: this cluster holds systems that have automated development and execution of a multi-step workflow, such as audience segmentation, journey design, or media buying. Like content creation, these usually let users specify a goal for the workflow and collect data to help measure results.

Customer Management: these products collect data to support optimization, help users to create messages, and select and deliver messages across customer touchpoints. They come the closest to the fully transformed state but don’t support all channels or create hyper-personalized messages in real time.

Transformed Products: This cluster would hold products that deliver the fully transformed process. There’s a good chance that some vendors have this mind, but we haven’t seen any products that deliver it. So the cluster is empty.

Infrastructure

Infrastructure and Business Models won’t take their final shapes until the transformed process is fully deployed, but they do co-evolve with the application changes. This applies especially to supporting infrastructure, parts of which must be in place for some intermediate product to function effectively.

Supporting infrastructures: The key supporting infrastructures for transformed customer management are inputs (data access and quality), internal processes (agent cooperation and analytics), and outputs (media integration). We see announcements in all these areas.

Data Access: most data access announcements describe accessing data in real time without loading it into a separate database. So far, these developments focus primarily on reading data in the company’s internal systems. But remember that the full vision for transformation includes access to third-party data such as compiled customer behaviors and local weather. We do see some of that, although none is in the current sample.

Data Quality: these tools prepare data for AI use. Many data quality vendors have added features to support AI use and have added AI-powered offerings. These apply to customer data but we don’t usually write about them, so this cluster is fairly sparse.

Agent Deployment: this describes technology for building agents and helping agents work together. It’s another field with extensive activity that is largely beyond the scope of the daily newsletter. It will be critically important if the transformed state involves teams of agents that cooperate closely with each other.

Analytics: any goal-seeking agent will need internal analytics to guide its decisions. Still, there may turn out to be a market for independent agents that make their results available as a shared resource for teams of specialist agents. This would apply especially to customer value analytics, which are needed to compare opportunities across different channels.

Hightouch Adds Journey Simulations

Media Integration: this is the output infrastructure that deliver messages created by the hyper-personalization. The tools are steadily encompassing more channels.

Surrounding Infrastructures: changes to the surrounding infrastructure happen after the transformed application is in place. This makes them harder to predict than supporting infrastructure, which develops sooner. We do have some current developments that are likely to become more important as hyper-personalization matures. There will no doubt be others.

Integrated Commerce (Sales Agents): this describes shopping in non-traditional channels, including retail media, search engines, social media, video, connected TV, and mobile apps. These can work without hyper-personalization but are vastly more effective when offers are tailored to the individual and context. In many cases, the interface will be a chat-style, AI-based sales agent that has a real-time dialog with the customer.

AI-Based Search (Buyer Agents): this describes marketing that is targeted at AI agents rather than humans. Today's most common implementation is AI search overviews, which do research for buyers. This changes the goal of search marketing from attracting traffic to appearing in genAI summaries. Other agents are evolving for other types of research and other stages in the sales cycle including making purchases on the customer's behalf.

Interacting with these buyer agents at scale and cost-effectively will require AI-based sales agents. In some ways, it won’t matter whether the sales agent is interacting with a person or a bot: to the sales agent, both are collections of data that must be analyzed and responded to appropriately. That said, the behaviors of humans and bots will be significantly different, so the sales bots will no doubt develop separate approaches to each. Ultimately, marketing systems may give buyer agents direct access to (some of) the product and promotional information they provide to sales agents, bypassing the sales agents altogether.

Because this is a relatively new development, our news coverage includes more research reports than product announcements.

Business Models

Few companies make announcements about changes in their business model, especially when those changes involve firing people. As a result, we have few newsletter items on AI-based business models. One change we do see is movement towards pricing based on value created rather than resources consumed. The transformed process will surely create other opportunities, perhaps including “no staff” companies run almost entirely by AI or “no product” companies that source products in real time as they find interested customers. It's likely the new business models will include ones we haven't even imagined.

What We’ve Learned

The main purpose of this blog post is to determine whether the STIB model provides a useful way to think about technology innovations. Of course I’m biased, but I believe it does. Not only was it reasonably easy to slot products into different categories, but the process generated several helpful insights:

Products that substitute AI for one step within an existing workflow are significantly different from products change the workflow.

It helps to envision the final, fully transformed product so we can assess intermediate products, identify the supporting capabilities and infrastructures it requires, and predict the surrounding infrastructure and business models it is likely to generate.

We should acknowledge that our prediction could be wrong. It may help to build scenarios around alternative outcomes.

Intermediate products appear before the final transformation. These become possible as new capabilities, infrastructures and business models appear. Observing the intermediate products helps to assess whether the changes are moving in the direction we expect or to adjust our prediction of the transformed state..

A minimum set of capabilities, infrastructures, and business models must be in place before the final design can succeed. Companies will fail if they offer the transformed design before supporting infrastructures are in place.

Capabilities, infrastructure, and business models continue to evolve after the (successful) transformed design is introduced. These developments make the product more effective and exploit the opportunities it creates.

Continued evolution creates rapid growth and expands the value of the final design, giving it an increasing advantage over alternatives. This advantage ultimately locks other configurations out of the market, even though some may be technically superior.

Applying the STIB model to customer management offers additional insights. Our sample of product news is enough to show:

Many current products do simple substitution. These are the easiest to deploy and can show immediate improvements in cost and quality. (There are also many products that support the existing workflow without any changes, but those don’t show up in our sample.)

Some intermediate products are already available. Most of these automate a single stream of tasks within the customer management workflow, such as content creation, campaign design, media buying or analytics.

Intermediate products usually work within a single department. This makes them easier to drop into the larger workflow and reduces the number of users whose work is disrupted.

Some products aim to automate an entire workflow, such as campaign design, development, and execution. Doing this with autonomous AI agents is the leading edge of the industry today.

Some companies offer components that support the final vision. These include application capabilities such as content optimization, response simulation, and advanced attribution, as well as infrastructure and business model changes including agent coordination, cross-company data sharing, touchpoint integration, and performance-based pricing.

I haven’t seen any products that promise the predicted final state of “real-time, hyper-personalized, omni-channel messages.” This has surely occurred to many smart people, so I’ll guess they have decided (correctly) that the capabilities, infrastructure and business models aren’t ready yet.

The final state may be delivered by a single integrated product or by multiple agents working in concert. Remember that the final state requires close cooperation between advertisers and media companies and between different departments within the same company. This makes a multi-agent solution more likely and reinforces the needed for data- and process-sharing technology..

Implications

Here are some practical implications of what we’ve discussed.

Substitution products can be valuable, but the benefit won’t last. It’s tempting to argue that substitution is a poor investment because it offers only incremental improvements on today’s current processes and that companies, and software developers should instead focus on more profound solutions. But the plain fact is that substitution creates substantial benefits and is easier to deploy than changes that require process change. Product developers and business users shouldn’t shy away from substitution but they do need to recognize it has a relatively short shelf life. Product developers should also realize that nearly all incumbent vendors will add substitutions to their current products or have already done this. That makes it hard to convince users to change to a new system on the basis of substitution alone. Attracting new clients will require solutions with more substantial advantages, which is what the intermediate product designs can offer.
The goal of "real-time, hyper-personalized, omni-channel messaging" is not yet widely discussed. What leads me to expect this as the transformed state is the growth of direct sales in social media, search results, connected TV, podcasts, and pretty much every other channel. This "instant commerce" collapses the multi-step "awareness, interest, desire, act" cycle into a single moment when the customer is presented with an opportunity to buy. Taking full advantage of this moment requires marketers to connect with all message opportunities, to gather all data so they can assess the potential value of each opportunity, and to deliver the most effective possible message for each opportunity they purchase. Only AI can do this effectively at scale, most likely by presenting intelligent agents to interact with customers who engage.**
The transition can be gradual. Instant commerce can be deployed in one channel at a time, can work with limited data, and doesn’t require advanced message optimization. Companies and vendors can move towards the fully transformed state in stages, building experience, product, infrastructure, and business models along the way.
Instant commerce depends on long-term relationships. That may seem like a paradox but customers will only engage with companies that they trust. There’s no time to build trust during the interaction moment, so trust must be built in advance. The good news is that the importance of trust is widely recognized and the methods for building trust (among humans) are well understood, if not always well executed. The industry may need new lessons in building trust among AI agents.
Infrastructure may offer the greatest opportunities. The biggest gaps between what’s currently available and what’s needed in the transformed state seem to be deep connections to collect external data and interact with touchpoint systems.

Shallow connections already exist, but real-time, hyper-personalized, omni-channel messaging implies real-time queries of external data sources about individual customers to collect up-to-moment information on behaviors. The touchpoints where those behaviors occur must capture, identify, assess, expose, and charge for that data in real time in a privacy-compliant fashion. Bear in mind I’m talking about touchpoints outside the company that wants to use this data. Very little technology exists to do that today. Data clean rooms are a start.
Beyond sharing information, those same touchpoints need to receive ad messages and deliver them to their visitors, again in real time and with feedback on response. Most of the messaging technology will reside with the ad buyer or middlemen, who will need to receive notification of contact opportunities, gather data and assess those opportunities, select the appropriate message, bid on delivering the message, transmit the message on bids they win, and measure results. Again, the existing technology to do all this in real time, at scale, and across many channels is limited at best. (Today’s programmatic ad system is a partial model.) In addition to data movement, this process requires sophisticated evaluation models so marketers can accurately bid on the projected value of each interaction.
Because all this interaction is happening between different companies, the infrastructure technology must be widely shared. This could imply broadly accepted standards implemented by many developers, or, more likely, proprietary technology built and sold by a few major suppliers. Competition to be one of those suppliers will be fierce but the rewards are likely to be huge. The rewards might diminish over time once the process is well enough understood to create open standards that are a viable, cheaper alternative.

Data quality doesn’t get the attention it deserves. Survey after survey shows that data issues are the top roadblocks to marketing, personalization, and AI success. Yet companies rarely make data quality an investment priority. There isn’t much to say about this except that real-time, hyper-personalized, omni-channel messages make data quality more important than ever. This is yet another piece of infrastructure that’s ripe for improvement.
Buyer bots might change everything. Buyers have limited attention, which is why gaining attention has always been the first step in successful marketing. This still applies to real-time, hyper-personalized, omni-channel messages -- so long as they're being sent to humans. But if customers delegate their purchasing activities to bots, the fundamental truth is no longer true: buyer attention will no longer be limited. Developments such as search engine optimization and AI search overviews as early examples of marketing to bots: marketers must target the algorithms, not attract human attention. But search marketing of any kind is still aimed at putting messages in front of eyeballs. Truly automated purchasing will remove humans from the entire process. The change won't happen overnight but it seems plausible to expect a mix of human and bot buyers in the near future. A STIB model with bot buyers as the final transformed state would be quite different from the one I’ve presented here.

Summary

This post has explored whether current developments in customer management technology can be effectively analyzed with the STIB model of innovation diffusion and whether the results provide useful insights into industry trends. I believe the answers are Yes and Yes. I've also presented a specific vision for the industry future, of "instant commerce" delivered through real-time, hyper-personalized, omni-channel messages. I can't promise that this is correct, but it's an interesting starting point for discussion.

_________________________________________________________________________________

* The graphic is conceptual but could be quantified by counting the number of components in each product that match the final transformed design or that differ from the initial state. Fun!

** Things are admittedly a bit more complicated for non-impulse purchases. But I’d argue you’re still trying to motivate an action in the moment, even if it’s only saving an offer in a wallet for future consideration.

Predicting the Future of AI-Based Customer Management: Lessons from Machine Tool Motors

2025-05-11T10:58:00.004-04:00

As someone who has spent his career tracking new technologies, I’ve long been fascinated by how those technologies spread over time. The standard model of technology diffusion, developed by Everett Rogers and based on New Deal studies of how farmers adopted new crops, offers the familiar segmentation of users into innovators, early adapters, early majority, late majority, and laggards.

It’s probably best known in the tech world in its Crossing the Chasm incarnation, which adds the all-important chasm between the second and third stages.

But there’s another model I find more helpful. This looks not at who adopts new technologies but at how they’re deployed over time. There's a consistent pattern:

Substitution (the new technology is deployed to exactly replace the old technology without changing the surrounding processes or workflows)
Transformation (the process is adapted to take advantage of the capabilities of the new technology)
Infrastructure change (new infrastructure is developed to support the transformed process)
Business model change (a new business model emerges to take advantage of the transformed processes and infrastructure)

I haven’t found any single source that presents this model in quite these terms, which may mean I could name it after myself, although for now I'll just use STIB. Some close matches include Ruben Puentedura’s SAMR Model for educational technology (substitution, augmentation, modification, redefinition), Deloitte’s Three Horizons Model (process optimization, process flow and quality, new business models) and Carlota Perez’s theory of technological revolutions (radical innovation, initial optimization, incremental innovations, maturity).

The paradigmatic example, at least for me, is deployment of electric power in U.S. factories. This started with connecting electric motors to the shaftworks that were previously driven by waterwheels or steam engines (substitution). After several intermediate changes, the endpoint was attaching a separate motor directly to each machine tool (transformation). This enabled an infrastructure change: no longer constrained by the limits of mechanical power transfer or proximity to water or coal, factories became cleaner, larger, single story, and located near other resources or markets. The new infrastructure resulted in a new business model featuring centralized mass production and standardized national brands.

Perhaps you can't relate to machine tools. Fair enough. Consider the automobile instead. The first cars were literally horseless carriages, pretty much the same as horse-drawn carriages except that they used a mechanical motor instead of a horse. That was substitution. Over time, the design of the cars was modified to take better advantage of having an internal combustion engine. That was transformation. Once it became clear the new model was going to be a success, manufacturers developed new, mass-production business models that better suited the new products and processes. Ultimately, this was supported by a new infrastructures of roads, filling stations, repair shops, traffic rules, licensing requirements, insurance products, and ultimately an entire, auto-based suburban landscape.

This may seem academic, but it applies directly to something you probably care about: the growth of AI. Seen through the STIB framework, most of the AI applications we see today are substitutions: an AI copywriter replaces a human copywriter in an unchanged workflow. The omnipresent co-pilots are another, even less disruptive type of substitution: they help humans perform the same tasks more efficiently, again without changing the workflow.

But it’s clear that these substitutions don’t take full advantage of AI’s potential. Again referring to the STIB framework, the question is what the transformed applications will look like. To answer this, we conduct a thought experiment: what would the process look like if it were redesigned to make best use the new technology? For electric motors in factories, the final form was motors attached directly to machine tools. For most customer data and customer-facing operations, the final form of AI is likely to be a single process that completes all the previously-separate steps at once. This is because AI is not limited by the need to have different specialists perform each step in the workflow, a constraint which results from the inability of mere humans to master more than one specialty, and from the need for experts to check the output of each step before moving on to the next.

But I don't think unified execution is the final form of the AI transformation. A deeper change is likely to remove discrete units such as customer segments, content pieces, campaign flows, and maybe even standardized products. Those exist because humans could only manage small numbers of segments, messages, campaigns, and products. An AI could handle more-or-less infinite numbers of these, which in practice would mean treating each customer and each interaction individually.

This leads to an end-state of “hyper-personalized” messaging, where content is custom generated on the fly for each customer and context.

Imagine an all-knowing, all-seeing bot that listens to what’s happening in the market and jumps into action each time it sees an opportunity to do something useful. The action will be optimized using all relevant data, including the company’s own information about the customer; second-party, third-party, and public information about the customer; behaviors of other customers; and market conditions, inventory, and who knows what else. In another dimension, this listening can extend beyond company-owned systems such as websites and contact centers, to include appearance of customers on third-party sites (already available to some degree through programmatic ad bidding) and even in walled gardens (which already receive lists of customers to watch for; the change would be to open a channel that lets the company assess the situation, generate the optimal message, and send it back for the walled garden for delivery.)*

It should immediately be clear that this vision requires infrastructure and business model changes from what’s available today. A much-improved data sharing infrastructure is needed to monitor behavior and access data outside of company-owned systems. This implies new business models to compensate external data owners for access to their information. Perhaps the data owners would charge a fee for letting companies monitor their data streams or query their data stores; or maybe they would only charge for data that a company uses; or perhaps the fees will be based on outcomes such as clicks or sales. Most of these schemes will ultimately require some way to estimate the value contributed by a particular piece of data.

Hyper-personalized message delivery requires more infrastructure and business model innovations. One of the most important developments in marketing today is the emergence of new channels that allow direct customer interaction: these include interactive TV, social commerce, online games, retail media, and even interactive podcasts and out-of-home advertising. All are alternatives to common web display ads, social media ads, and search ads, which are also becoming more interactive. As with data, the key change made possible by AI is the ability to monitor vastly more opportunities at once, to evaluate the potential of each opportunity in real time, and to take advantage of the opportunities offering the greatest value.

I certainly hope that everyone reading this realizes that what I’m describing is far beyond the capabilities of today’s AI systems. The data access process requires AI to continuously ingest, clean, and integrate data from multiple sources and to automatically adapt as new sources appear and established sources change. Remember that AI is just beginning to address the bottleneck of incorporating new data sources into today’s CDP and warehouse systems. Similarly, we’re just beginning to see AI systems deliver intermediate steps on the way to hyper-personalized messaging. Today’s cutting edge is automated campaign design, which at best (and with much-needed human quality checks) could transform a user’s prompt into a complete campaign package of audience selection, content, and delivery rules and then execute that package. While impressive, this still uses the conventional structure of a few, discrete segments, content pieces, and campaign flows. That makes it closer to substitution than true transformation.

Another way to look at this is that the vision offers a roadmap for future AI development. The current frontiers in AI are goal-seeking agents, access to external data (Anthropic’s Model Context Protocol), and agent cooperation (Google’s Agent-to-Agent). (See this Medium post for a good overview of these.) If my vision is correct (which is by no means certain), steps beyond those frontiers will include proactive data gathering and integration, automated data value assessments, greater situation awareness, and better simulation of human behaviors. (I'd really like to say "understanding" of human behaviors but don't think we can quite attribute that to AI.) AI will also need more economical processing and reliable guardrails against hallucinations, biases, privacy breaches, and generally bad behavior.

The table below offers a more detailed view of where I think things are headed. It looks at four major customer data processes: customer data management, people issues related to customer data management, customer data activation, and advertising. For each process, it lists the required capabilities for each of the four diffusion stages. You can think of these as requirements for new, AI-based products.

AI Applications for Customer Data (STIB Model)
	Data Management	People	Activation	Advertising
Substitute [execute via co-pilots and agents]	Data collection, ID resolution, connectors, metadata	Understand applications, requirements, training	Segmentation, analytics, prediction, sharing, privacy	Audience assembly, media buys, data buys
Transform [execute via unified AI systems]	Unified process, data as service	Management tools, define goals/prompts, explore opportunities	Hyper-personalize messages	Deliver best customer, data, channel/media; optimize spend
Infrastructure [required capabilities; many delivered via AI]	Automated data access, security, privacy, quality, transforms	Learning systems, training systems, process design systems	Efficient processing, attribution, instant commerce, buyer agents	Secure data sharing, consented data assembly, contextual targeting, marketplaces, fractional billing
Business Model [rely on AI for analytics and operations]	Value-based pricing	Pay for skill achievement	Goal achievement, Sales as a service	Goal achievement, audience as a service, value-based pricing

Where does all this lead? Here are the main points I hope you’ll take away:

1. The impact of AI on customer data is just beginning. We can expect AI to be deployed in the same pattern as other technologies. At first, it will substitute for humans or non-AI systems by performing the same tasks within existing workflows. Over time, it will transform those workflows into new processes that take full advantage of what AI can do. Ultimately, the industry will develop new business models and infrastructures to support the transformed processes.

2. For developers: consider which stage your AI project is targeting. While substitution is low-hanging fruit, bear in mind that current processes will soon be obsolete. Consider developing products that support transformed processes and their related business models and infrastructures.

3. For users: current self-service AI tools may enable you to build your own substitutions with minimal investment. Unlike system developers, you can afford to deploy these now and then discard them when something better comes along.

4. If you have a vision for a transformed process, you could try to build it. It won't achieve its full potential because the supporting AI capabilities, infrastructure, and business models are not yet available. But it might deliver enough value to justify the investment: imagine hyper-personalization based only on one company's own data and deployed only on the company's own customer-facing systems.

5. The final form of the transformed process will emerge over time. As companies experiment with different approaches, the industry will converge on an optimal design. Commercial developers will then build systems with this design and flesh out the supporting AI features, infrastructures and business models. As with most complex systems, commercial vendors will probably own most of the market because they can afford to invest more in their products than most individual firms.

6. Standardized integration mechanisms will play an important role in this new world. This assumes I'm right that the transformed process will rely heavily on connecting to external data and external delivery channels. This should make integration a particularly fruitful area for investment if you're a developer, or for building expertise if you're a user.

__________________________________________________________________________________

*As a practical matter, it's unlikely that thousands of companies would separately monitor thousands of data sources and delivery touchpoints. It's more likely that data sources and delivery systems will broadcast event information which marketers can read and react to as they see fit. Perhaps the sources and touchpoints will package their events into channels covering different customer groups or event types and let marketers subscribe to the channels they want.

State of Martech Report: Customer Data Platforms Are Evolving, Not Dying

2025-05-06T21:27:00.012-04:00

Scott Brinker’s State of Martech report has grown from a one-page logo jigsaw to a small industry, co-authored with MartechTribe’s Frans Riemersma and apparently produced by a small army of elves reviewing thousands of products each year. The latest edition, released yesterday, provides the now-expected deep analysis of industry trends covering category growth, the impact(s) of AI, stack architectures, and more. It’s all interesting and too complicated (or maybe complex?) to recap here, so I can only suggest that you read it on your own.

In addition to data based on the 15,384 vendors now listed (up 9% from 2024), the report analyzes results from a survey of martech and marketing operations leaders. Again, there’s lots of fascinating information, but I was of course drawn to the sections related to Customer Data Platforms.

On its face, the report offers some pretty bad news for the CDP industry: the fraction of companies citing a CDP as the “center” of their martech stack fell from 15.5% in their similar 2024 survey to 12.5% in 2025. However, the 2025 survey is based on just 96 responses, meaning that’s a swing of three answers, so it’s not cause for too much alarm.* Still, it’s an interesting result, and even more intriguing when you separate B2B respondents (52% of the sample) from others (14% B2C and 34% mixed B2B/B2C). CDPs have never been widely adopted in the B2B world, and indeed, their share is unchanged from 2024 (7.9%) to 2025 (8.0%). The flip side to that is in the B2C and mixed group, the shift is larger: from 26.9% in 2024 to 17.4% in 2025. But, again, bear in mind the sample size: that 17.4% represents seven responses. A shift of three answers is well within the range of statistical noise.

Let’s put sampling issues aside and assume there’s some drop-off in reliance on CDP. The question is, what has taken its place?

Did you just answer “cloud data warehouse, of course”? That’s not entirely wrong – the data warehouse share grew from 20.9% to 23.9%. But the big winner was MAP/CEP (marketing automation/customer experience platform), which grew from 19.4% to 26.1%. CRM grew from 17.9% to 19.5%, or nearly as much as data warehouses. Multi-product suites fell from 1.4% to zero, which hints quite strongly that the respondents heavily skewed away from large enterprises.**

If we combine the MAP/CEP, CRM, and DXP or ecommerce categories into “customer-facing systems”, the combined share of that group grew from 43.3% to 52.1%. So if I were to read any trend from this data, it would be that companies are centering their martech stacks on customer-facing systems, not on data warehouses.

This is actually consistent with the trends we’ve seen in the CDP industry itself, where the most recent major acquisitions (ActionIQ by Uniphore , Lytics by ContentStack, mParticle by Rokt) all involved merging a CDP into a customer-facing product, and where customer-facing vendors like MessageGears, Klayvio, Insider, Listrak, and Braze have added CDP (or CDP-ish) capabilities***. The CDP Institute classifies all of these systems as CDPs, in addition to whatever else they do. So, the way I see it, companies that list those products as the “center” of their martech stack are still building their stack on a CDP, even if they don’t call it one. That’s good news for the industry, not bad.

The survey also takes another look at the role of data warehouses in the martech stack, asking “Do you have a customer data warehouse/lakehouse integrated with your martech stack?” More than half the respondents (56.2%) said they did, a number that climbs to 92% for B2C vendors and 80% for enterprise companies. What’s interesting here is the relation of those answers to the previous question: many more firms have integrated a warehouse than are using their warehouse as the center of their martech stack. This is far from shocking but, again, suggests that the mere presence of a warehouse doesn’t mean that warehouse is the primary customer data store.

If you’re a “warehouse-centric” composable CDP vendor, you could read the relatively small share of warehouse-as-central system either as cause for alarm – the market isn’t as big as you thought – or reason to rejoice – the potential for growth is huge. Both could be true at the same time. But if the primary industry trend is for CDP functions to migrate to customer-facing systems (run by marketing or other business units), then a shift of CDP functions toward the (IT-controlled) data warehouse seems to be the wrong way to bet. (In this context, the recent sale of leading composable CDP Census to data movement platform Fivetran may signal incipient consolidation in the young-but-already-overcrowded composable CDP sector. That Census was bought by a data movement tool rather than a customer-facing system reinforces the notion that composable CDP products serve IT teams while customer-facing CDPs serve marketers and other end-users.)

Presumably all those non-central warehouses are acting as data sources to CDPs or customer-facing systems with a CDP inside. Indeed, the Brinker/Riemersma report positions CDPs (stand-alone or embedded) as intermediaries between data systems and activation systems (which they call "systems of knowledge" and "systems of context," respectively), responsible for "organizing or framing the data in a way that best serves more situational needs." I quite agree, and couldn't have said it better. I would expand on this by noting that CDPs within customer-facing systems will have direct access to the data those systems generate, so the role of the data warehouse is limited to supplying data that the warehouse collects elsewhere. A CDP with direct access to customer-facing systems overcomes one of the main drawbacks of the warehouse-centric approach, which is that warehouses often can't provide the real-time access to behavior data that's needed for many customer-facing applications of CDP data.

I should stress that CDPs are a bit player in the Brinker/ Riemersma epic. Definitely download the report (it’s free and ungated) and focus on whichever portions you find most relevant. It’s all good.

------------------------------------------------------------------------------------------------------------------------------------

*The 2024 survey had 168 responses, of which 60% were B2B, 9% were B2C and 31% were mixed. Doing that math, that’s 67 responses in the B2C and mixed group, of which 18 cited CDP as the center of their stack.

**The 2025 report says 30 of the 96 responded were from ‘enterprise’ organizations, but doesn’t indicate how these split between B2B, B2C and mixed.

***In addition, at least one unacquired CDP (BlueConic) has repositioned as a customer-facing system.

Rokt Buys mParticle

2025-01-16T23:13:00.003-05:00

mParticle announced yesterday it is being acquired by ecommerce offer personalization vendor Rokt for $300 million. It’s the third acquisition of a leading CDP in a little over one month, following ActionIQ/Uniphore and Lytics/ContentStack. It’s also the first to announce the deal price.

I’ve already written in this blog about the other two deals and this one is similar enough that most of what I said before still applies. It’s clearly tough to be an independent CDP these days. ActionIQ, Lytics, and mParticle are all technically advanced products, with a particular stress on delivering real-time data – something that’s hard for cloud data warehouses like Snowflake and Google BigQuery. This is probably what made them attractive to their buyers, which are all in different businesses (conversation management, digital experience management, and ecommerce, respectively) but share the need to manage real-time customer interactions.

The buyers' businesses themselves increasingly overlap, so part of their motivation is likely to differentiate themselves from competitors. Probably more compelling, a CDPlets them position their product as the foundation platform in their clients’ customer management architecture. The platform is a powerful position because it’s tough to change platforms; applications that read data from a platform are precarious because switching applications is relatively easy. (And I do mean relatively: it’s still a lot of work in most cases.) On a more practical level, adding a real-time CDP makes it easier to do personalization, which offers a concrete advantage to potential buyers of these products. (And, yes, the data is essential for Artificial Intelligence as well.)

It’s interesting that all three buyers are relatively small firms (Uniphore has 851 employees, down 3% in the past two years; ContentStack has 609 employees, up 27% in the past two years; Rokt has 614 employees, up 21% in the past two years) and none are the biggest in their space. While they’re not struggling to survive, they do need a way to distinguish themselves from their many competitors. This is provided by the comprehensive data assembled by a CDP and the superior personalization it makes possible.

From the CDP vendors’ viewpoint, becoming part of a customer-facing system lets them reverse the trend of selling to IT and data departments and move back to the original CDP position of selling to business users. This is a more comfortable environment, since IT and data teams are much more likely than business users to want to build their own solution based on a cloud data warehouse. It also removes some competition from the giant enterprise cloud vendors like Salesforce, Adobe, and Oracle. Those companies are most appealing to central IT and data teams, although they certainly sell to business users as well.

There’s nothing radical about a CDP being embedded in a customer-facing system. In fact, statistics in our Industry Update report have long shown that what we call “campaign” and “delivery” CDPs comprise more than two-thirds of the industry. The truth is, the market long ago decided it preferred a CDP that was part of a larger product. So the latest round of acquisitions reflect a continuation of that situation, not a radical departure.

The obvious question is whether more acquisitions will follow. Likely candidates are firms that, like ActionQI, Lytics, and mParticle, have particularly advanced real-time technologies. This isn’t everyone, since many CDPs rely on standard database technology and differentiate in other ways such as industry expertise or regional presence. The biggest remaining independents (Tealium at 599 employees*, Optimove at 504 employees, and Treasure Data at 484 employees) might all be large enough to continue going it alone or at least might be too big for companies similar to Uniphore, ContentStack and Rokt to swallow. The next level down would be firms including Twilio Segment (a perpetual acquisition candidate as a spin-off from the rest of Twilio), Resulticks (293 employees), and Cordial (245 employees). Note they are still larger than mParticle (237 employees), ActionIQ (152 employees) and Lytics (49 employees).

More likely candidates are firms including BlueConic (177 employees), Redpoint Global (157 employees), and FirstHive (114 employees), as well as a variety of still-smaller firms including Blueshift, Simon Data, Lemnisk, Lexer, Commanders Act, Meiro, NGDATA, and Relay42. Several of this latter group are based outside the U.S., which may reduce their appeal, and of course every company has a unique situation. So it’s not clear which would actually be open to being acquired or be technically interesting to a potential buyer. But it wouldn’t at all be surprising to see one or more of these join the list of CDPs embedded in customer-facing system.

________________________________________________________________________________________________________________________- *all data from LinkedIn, as captured in the CDP Institute’s latest Industry Update, available here.

CDP Lytics Bought By DXP Customerstack

2025-01-09T07:15:00.004-05:00

Almost one month to the day since ActionIQ was bought by Uniphore, another first-generation CDP has been sold to a large customer experience vendor. This time it's Lytics, founded in 2013, which announced this week that it had been purchased in December by composable CMS and digital experience platform Contentstack.

Like ActionIQ, Lytics is a technically advanced product that has kept up with industry trends, making its features available as components, stressing real-time capabilities, and offering a broad range of campaign management and personalization features. In fact, just last October, Lytics announced integrations with major content management systems and a free self-service offering that built customer profiles and used them to make personalization recommendations available within any content management system.

But Lytics never achieved the scale needed to support its vision. The company raised a relatively meager $58 million, with its last round in 2019. (By contrast, ActionIQ raised $145 million including a $77 million round in 2021.) Headcount (per LinkedIn) fell from a peak of just over 200 in 2020 to 49 this January.

The Contentstack combination will provide the resources needed to continue developing Lytics’ vision. Contentstack has nearly 600 employees, up 24% in the past two years, and $169 million in funding, with their last raise of $80 million in 2022. They also claim over 500 customers, many of whom can be expected to adopt Lytics technology.

Indeed, the architectural fit between Lytics and Contentstack appears to be excellent. Both offer composable products, which should make it easy to integrate Lytics profiles with Contentstack’s personalization engine. This is exactly the ability that the company touts in its announcement of the acquisition. While I can’t speak to the technical compatibility between the two systems, the fact that Lytics already had integrations with Drupal and Wordpress suggests exposing Lytics data to Contentstack will be fairly easy.

And what does the deal say about the larger CDP industry?

One observation is that the adoption of a composable approach was not enough to preserve the independence of either Lytics or ActionIQ. Both those firms were early leaders in offering their systems as components. Maybe this generated some revenue but it was clearly not enough to turn around their businesses. I’d guess the price tags for components are too small to really help much, especially compared with the much larger fees these companies were charging for their integrated products.

The large number of competitors in the component marketplace probably ensures that prices will stay low, which is good for buyers but not so good for vendors. This doesn’t bode well for “composable CDP” vendors who only sell components, although presumably they have geared their organizations to be profitable while selling relatively low-ticket items. Note that “composable CDP” vendors like Hightouch and Census have all added multiple components, enabling them to earn more money from each client. But so long as competition keeps down the price of individual components, the revenue per client will remain low even from clients who buy all the components those vendors have to sell.

It’s likely that the investments that Lytics and ActionIQ made in “componentizing” their products made them more attractive to their ultimate buyers. But, otherwise, those investments may have harmed their business by consuming resources that could otherwise have been spent on improving their core products.

A second question is whether these transactions mark the start of the industry consolidation that analysts have been predicting pretty much forever.

I think not. Actual consolidation would involve mergers between competing CDP vendors, something that has almost never happened. (Exceptions have been in Europe, where Spotler combined Datatrics and Squeezley, Splio acquired Tinyclues, and Easyence merged with mediarithmics . You’ll note these are not exactly household names.) What we have seen over the years is deals like Lytics/Contentstack, where a larger customer experience vendor buys a CDP to add profile building, and sometimes campaign management, capabilities. I do expect these sorts of acquisitions to continue, as the growth of personalization and AI makes the need for unified data more pressing. We’ll also presumably see a few of the small independent CDPs vanish entirely, although this too has been unusual: enough companies need profile-building technology that the intellectual property of even a struggling CDP can usually find a buyer. In this regard, it’s probably not coincidental that both Lytics and ActionIQ were quite advanced technically, making them particularly appealing.

So, while I don’t see a literal consolidation in the sense of CDP vendors merging with each other, I do expect to see a continued drop in the number of independent CDP vendors. These will be replaced by customer experience vendors who include a CDP inside their products, the better to support personalization. To be clear, what will qualify these firms to be considered CDPs is the ability to work with external systems as data sources for unified profiles and as consumers of those profiles. This data-sharing ability will become more important as the number of customer-facing channels continues to increase and as customers expect consistent, personalized treatments across all of those channels. Since no single system can expect to manage all channels by itself, it becomes essential even for customer experience products to capture data from external systems and to support personalization across those systems. That was, and remains, the defining feature of a CDP.

Uniphore Buys ActionIQ: What We Can Learn

2024-12-08T19:51:00.006-05:00

Acquisitions of, or by, major CDP vendors have been vanishingly rare in recent years. So news that ActionIQ had been purchased by Uniphore generated an especially large amount of interest. Let’s look at it from three perspectives.

1. Significance for ActionIQ. Although ActionIQ has been a significant CDP since its founding in 2014, the company was always a bit different from most other CDPs. Rather than starting as a tag manager or messaging platform, ActionIQ’s roots were in big data technology: co-founders Tasso Argyros and Nitay Joffe were, respectively, the founder of big data engine Aster Data (purchased by Teradata in 2011) and a leading data engineer at Facebook. Not surprisingly, the company’s approach to CDP was also driven by data technology: in particular, it used in-memory storage to support high volume, real time use cases. In recent years, the company evolved its technology to offer component-based and warehouse-native options in addition to a conventional integrated CDP.

The company’s business results were less stellar than its technology. It primarily targeted large enterprises that could most benefit from its sophisticated approach, and had reasonable success in that market. But, despite raising a hefty $144 million (per Crunchbase), growth stalled in recent years. Headcount, which stood at 208 in mid-2021 (per LinkedIn) has declined steadily since then and last week was 161. We don’t know whether sales also fell; it’s likely that the headcount reductions reflected preemptive belt-tightening as new funding became less available, rather than tracking an actual revenue decline.

Either way, it doesn’t seem that ActionIQ was on a path for substantial growth as an independent company. So a sale to a friendly buyer like Uniphore makes good sense. I’m told that the ActionIQ product itself will continue to be sold, supported, and developed, so it’s a good outcome for ActionIQ customers as well.

2. Significance for Uniphore. I’ll admit I wasn’t familiar with Uniphore before this deal, even though they are in fact a well-established, fairly large company (founded in 2008, 800+ employees, $621 million in funding). Their specialty has been speech-related customer experience tech, such as self-service, phone agent and sales assistance and conversational analytics. Like others in that industry, they are currently focusing on enhancing their products with AI to improve worker productivity and business results.

What’s more interesting is where they differ with the competitors, which is a greater focus on the data infrastructure needed to support AI. This is where the ActionIQ acquisition makes sense, since it will help to support what Uniphore calls the “Zero Data AI Cloud”. As the name suggests, Uniphore’s vision is to provide AI systems with data without first exporting that data into a separate database. This is the “zero copy” or “warehouse-native” approach that’s increasingly popular in CDP circles. It’s also something that ActionIQ has evolved to support, so the acquisition is strategically sound. As you’re probably aware, Uniphore also announced its acquisition of Infoworks at the same time as the ActionIQ acquisition. Infoworks is an even more technical vendor, dedicated to managing cloud data migrations. At 75 employees and $71 million in funding since it was founded in 2014, it's smaller than ActionIQ but still a substantial business. Like ActionIQ, its head count has also fallen in recent years, from a recent peak of 112 in early 2023.

Uniphore’s pivot to AI data infrastructure seems like a good move, at least compared with the hugely overcrowded AI-based CX marketplace. I think “pivot” is a fair term here, since Uniphore’s earlier acquisitions – Hexagone and RedBox in 2023, Colabor in 2023, and Emotion Research Labs in 2021 – all firms applied AI to more conventional CX use cases, such as emotion detection and conversation analysis. Of course, the data infrastructure space itself is the playground of major cloud companies like Google and Amazon, as well as cloud data firms like Snowflake and Databricks. Whether Uniphore can carve out a unique niche for itself as a tooling provider to connect those systems is uncertain but it seems worth a try.

As an aside -- I don't think a true "zero copy" approach to AI data is really possible. We had this debate in the early years of the CDP industry, when companies including Salesforce and Adobe argued they could assemble customer profiles on the fly without a persistent database. They ultimately learned that didn't work. This doesn't invalidate Uniphore's strategy, even though the actual implementation will almost surely involve extracting, cleaning, and reorganizing data and storing the results somewhere that AI systems can use it. That "somewhere" could in fact be a data warehouse -- the approach favored by composable CDP vendors. This might more accurately be called "one copy" than "zero copy," although it wouldn't sound as appealing.

3. Significance for the CDP Industry. ActionIQ’s decision to sell to Uniphore obviously says something about the prospects for independent CDP vendors, but the implications are limited because ActionIQ’s technology was atypical. In fact, since ActionIQ already supported the “composable” and “warehouse-native” approaches that are often touted as successors to traditional CDP systems, you could argue the deal indicates that prospects for firms taking those directions are more limited than their advocates believe.

Given ActionIQ’s unique situation, I don’t think this deal presages a sudden burst of exits by mid-tier CDP vendors. Even though growth has largely stalled for those firms, I think most can keep afloat long enough to find their footing in a new world. This footing may well differ for different firms: for example, we’ve seen Lytics attempt to simplify CDP deployment enough to make it accessible to smaller businesses, while mParticle is moving more in the direction of combining customer data with analytics. All need to dodge the giant cloud, cloud database, marketing cloud, and messaging vendors who are trampling the heart of traditional CDP territory. Some will indeed take shelter within larger organizations like Uniphore. Others may stay independent but retreat to niches such as data management tools or serving specific industries. So even though the market for traditional, integrated CDP products is likely to shrink, I believe most of the vendors will survive in some form or another. This is because the fundamental need driving CDPs – the need for unified, accessible customer data – will only grow stronger. Companies that find a profitable way to meet that need will be able to thrive over time.

Composable CDP is Dead

2024-10-29T20:54:00.006-04:00

Composable CDP is Dead. Sure, it’s a clickbait headline. But how else to gain attention on a topic that’s over-saturated with hype?

Of course, I would never say something untrue just to gain attention. So let me clarify exactly what’s deceased. It isn’t composability or CDP, or the vendors who have adopted the “composable CDP” label. What’s dead is “composable CDP” as a distinct category of software.

You might stop here to argue that “composable CDP” was never a meaningful category, and I wouldn't necessarily disagree. But the term did have a broad meaning along the lines of “software components that provide functions needed in a CDP.” What made it useful was the contrast with packaged (a.k.a. standard, conventional, traditional, integrated, or just plain “real”) CDPs, which, according to the CDP Institute definition, provide a complete set of CDP functions – specifically, the functions to collect, organize, and share customer profiles. Companies that wanted to build their own CDP, possibly using their existing data warehouse as a base, could use the “composable CDP” label to identify products that might help with their project.

What’s changed is that many of the packaged CDP vendors quickly caught on to the demand for CDP components and split their products into modules. This let companies buy only the functions they needed and apply those functions to data in a cloud data warehouse. The repackaging is easier for some vendors than others, depending on how their products were originally built. But enough of the original CDP vendors now offer some flavor of CDP components that companies looking for components to build an in-house CDP should consider them as well. Thus ends the utility of “composable CDP” as a category. RIP.

The composable CDP vendors saw this day coming. Many have switched their positioning to "warehouse-native," which is more accurate and reduces the focus on point solutions. But their main strategy has been to move beyond offering a single component for a single function – that is, their original product – to offer multiple components and a complete set of CDP functions.¹ The change to offering a complete solution weakens some of the composable vendors’ original sales arguments, including the superiority of “best of breed” solutions and the ease of swapping components from different vendors “like Lego blocks”. But it also allows new arguments, including the virtues of buying multiple preintegrated components from the same vendor and the value to end users of working with fewer tools. If any of the vendors feels embarrassed at switching from one side of the debate to the other, I haven’t noticed.

Then again, there’s no reason they should be ashamed. Products naturally expand their footprints: it’s a marketing truism that it’s easier to sell new products to existing customers than to sell existing products to new customers. And CDP buyers clearly prefer integrated solutions: according to the CDP Institute Industry Update, nearly three-quarters of CDP products extend beyond profile creation to include marketing execution. In practice, only very large companies can afford to build custom solutions by cobbling together “best of breed” products. I summarized this years before CDPs existed as “Raab’s Law”, which states that “Given a choice, most buyers prefer suites over best-of-breed solutions.” The pithier version is “Suites win.”

This doesn't mean that most companies actually use integrated marketing suites. The majority clearly do not. In fact, a recent Acxiom survey found just 17% of marketers plan to move to a suite in the next twelve months; the rest are using multiple products and 29% expect to add more. Reasons companies don't adopt suites include cost, agility, and the lack of required functions. Still, when you look at marketing software categories including campaign management, journey orchestration, analytics, and email execution, leading independent vendors have been acquired by suite owners and the remaining independents are largely relegated to niche segments or have expanded to become suites themselves. This is what the composable CDP vendors are doing.

Wait, did I just say that an integrated CDP is a software suite? To some degree, yes: in the finest Alice-in-Software-Land tradition, words mean whatever I want them to mean.² But the less flippant answer is that it depends on your perspective: in discussing CDP functionality, a single product that provides all necessary functions does look like a suite, while components for individual functions look like point solutions. But if you raise the perspective to marketing architecture, the CDP itself might be a single component alongside content management, marketing automation, email, and media buying components, while a marketing suite like Adobe would encompass them all. Go up to the enterprise level, and the marketing suite could be one component alongside finance, human resources, and operations, compared with an enterprise-wide suite like SAP.

It’s possible that Raab’s Law begins to weaken above the departmental level, where conflicting managers are powerful enough to insist on their own systems rather than accepting the loss of control implied by a shared system. Feel free to explore this on your own. What matters here is that talk of components reintroduces the topic of composability.

Although composable CDP only entered the martech discussion about two years ago, the broader label of composable architecture began to build momentum around 2020.³ Today it’s applied to nearly every type of software, and we see composable-oriented organizations like MACH Alliance expanding their scope accordingly.⁴

There’s actually a good chance that composability has hit its peak on the hype cycle, as greater experience gives practitioners a better understanding of its limits. As with composable CDP, these primarily involve integration costs and support complexity. They also include scaling challenges to move large volumes of data between large numbers of components, and problems when one weak component drags down performance across the entire structure. Call it the Revenge of Raab’s Law.

For composable CDP in particular, experience has also revealed high processing costs when everything is done within a cloud data warehouse. (This is why cloud data companies like Snowflake and AWS are so enthusiastic about the warehouse-native CDP approach.) Users have also discovered that the cloud data warehouses struggle with certain kinds of real-time processing, a core requirement for many CDP applications.

This more realistic understanding of composable limitations doesn’t mean that composability is dead. Marketers were never really all that interested to begin with: just 3% in the Acxiom survey said they plan to move to a composable strategy in the next 12 months, compared with 17% moving towards a suite.⁵ The people who really care about composability are the technicians: when Messagegears surveyed people interested in “data cleaning and quality control,” 56% said they prefer a composable architecture and 33% had no preference, leaving at most 11% in favor of packaged systems.

But don’t get too excited, composable fans: preference isn’t the same as intent, let alone action. You can bet many of those technicians won’t be able to make the changes they’d prefer.⁶Still, the technicians’ views matter a great deal because growing use of customer data beyond marketing gives IT and data teams a larger role in selecting customer data systems.

Will technicians get over their infatuation with composability? Only time will tell. But it’s likely their ardor will cool as they learn more about it and get distracted by whatever Big Thing comes Next.

This doesn’t mean the the pendulum will swing back to favor traditional, integrated CDPs. Rather, as Hegel would have predicted, we can expect a synthesis where most customer data is assembled and remains in the warehouse, but some functions are executed outside of the warehouse in a CDP that reads the warehouse data directly. In this world, traditional CDP functions such as data ingestion and identity management become components that operate within the warehouse, while functions such as real-time response, analytics, and journey orchestration become part of a separate, internally-integrated component that can reasonably be labeled a CDP.

The composable CDP is dead. Long live the CDP components.

_________________________________________________________________________________

1. The definition of “complete” varies from vendor to vendor, and inevitably matches whatever components the vendor has available at any particular moment. But that approach is common throughout the software industry, so it doesn’t’ represent uniquely deceptive behavior by the composable CDP camp.)

2. See the previous discussion of “complete.”

3. The idea of building systems from components has been around much longer than the label, using terms such as micro-services architecture and headless CMS. CDP itself is also much older than four years. All these terms are much more widely used than composable architecture and composable CDP.

4. I’ve never spoken with anyone at MACH Alliance, but surely they face the same struggles maintain consistent definitions and standards as the CDP Institute. Fun fact: while several of the traditional CDP vendors have met the rigorous standards required to join the MACH Alliance, none of the composable CDP vendors are members.

5. The remainder intend to muddle through with their current crazy wall of loosely connected applications, while complaining about integration cost and feature overlap.)

6. How big is gap between preference and planning? Well, in the Acxiom survey where 3% of respondents said they plan moving to a composable approach, 36% said they thought a “modern sophisticated marketing stack” should have a composable architecture. Let’s call them composable-curious. Of the rest, 38% cited an integrated marketing suite, compared with 17% planning to move in that direction, and 34% chose a best of breed approach, compared with about 80% living that particular dream.

CDP Round Table: Forget Composable, Are Cloud Databases a Threat to CDPs?

2024-07-18T11:20:00.001-04:00

On July 16 and July 18 2024, the CDP Institute hosted a pair of online roundtable discussions for industry vendors. Here are summaries of the conversations.

July 16 Roundtable - U.S. and Europe participants.

We started with a quick list of trends that the Institute is watching. These include Composable CDP, CDP integration with advertising channels, third-party cookie deprecation, vertical (industry-specific) CDPs, and AI applications in CDP systems.

The discussion quickly focused on Composable CDP. Key points included:

Replacing the term “composable” with “warehouse native”. This is more accurate and resonates in particular with data and IT teams, who want to get the most use from their warehouse investments. “Composable” is still more commonly understood among marketing users.
Composable CDP comes up most often at large enterprises, which have the most mature IT and data resources.
Composable also has appeal in healthcare and financial services, where regulatory concerns make companies reluctant to copy data into a separate CDP database. (Note: I have also heard the opposite, because IT and data teams don’t want to give business users direct access to their entire warehouse and would rather provide them with a limited extract.)
Digital native companies are also good prospects for Composable because they tend to have well managed data already in place.
Composable promises a faster start than building a separate CDP, which is appealing to all potential users. Whether this really happens depends on the state of existing data sources and warehouses.
Composable CPDs have an advantage because the company’s existing warehouse will include both customer and non-customer data elements tailored to company and its industry. Standard CDPs often must build a custom data model for each new industry, and may struggle to include non-customer information such as product data. Vertical CDPs are appealing because they have industry-specific data models already in place.

Other observations:

The fast growth of cloud databases, and Snowflake in particular, is a threat to packaged CDPs because they make it easier for IT to build their own products. So far, the cloud databases haven’t built in key tools such as data quality and ID resolution, but there are indications that will change. Such tools are already available as pre-integrated applications in the cloud database vendors’ marketplaces. This is worth calling out as a separate trend from Composable CDP.
The data engineers who buy cloud databases are not aware that CDP systems exist. They see the value when it’s explained to them, but are still likely to want to expand use of their warehouse systems to justify the investment. This is true even if the warehouse doesn’t already have rich customer data features and customer data has been a low priority. “Data-in-place” and “zero-copy” are strong selling points against putting data into a separate CDP.
Regulatory changes, such as anti-redlining rules for banks and Sunshine laws for healthcare marketing, have driven investments in customer data in the past. Privacy regulations may play a similar role in the nature future. This is also worth calling out as a trend.
There is limited convergence between CDPs and privacy systems. Few CDP vendors have invested in privacy features beyond ingesting consent data, possibly because privacy is complicated so it would take a major investment. In addition, privacy systems have different buyers from CDPs, and privacy managers feel it’s safer to buy from OneTrust than lesser known vendors. But basic privacy and security always requirements in CDP RFPs.
Journey orchestration is available in many CDPs, which means they overlap with existing journey orchestration systems. But very few clients will replace a mature journey orchestration system with a CDP, due to the effort involved in training staff and migrating programs. Most CDPs with extensive journey orchestration and messaging capabilities are vendors that started as journey orchestration and messaging products and added a CDP capability. Journey orchestration and messaging systems that connect directly with a warehouse may pose another threat to CDPs.
Advertising integration is often the first CDP application in Italy and elsewhere in Europe.

July 18 Roundtable - APAC participants

Industry trends

Overview: composable, cloud database, advertising integration, cookie deprecation, privacy & compliance, vertical industry CDPs, AI
Seeing lots of requests for clean room and consent management, and how CDPs can merge those to streamline work for customers. Some parts of consent can be managed in CDP, others should be in separate platform. Privacy is often first use case.
Among companies with CDP already deployed, often see extension beyond marketing to other departments. CDP often leads to teams within marketing working together that before kept separate.
When a new CDP is deployed, people become aware of it organically and through analytic reporting that uses CDP data.
Cloud databases and composable CDPs introduce new buyer personas from IT and CTOs, who have different use cases and technical concerns from marketers.

Other items

Understanding of CDP is relatively low in SouthEast Asia (SEA), compared with Australia, India and Japan.
Composable is mostly raised by IT teams, who are looking for easiest path to moving data from one platform to another, without understanding the strategy behind the CDP project. Applies to both small and large organizations. Companies like Salesforce with a lot of installed products can co-exist with composable.
Companies can easily identify many use cases for a CDP. Vendors spend time helping them to prioritize. The sales process is sometimes slowed because buying teams are overwhelmed by the number of use cases.
Greatest interest among marketers is performance marketing use cases, where results can be directly measured, such as data activation, segmentation, and pipeline-to-paid. This is limiting because there are so many other use cases that don’t have immediate results, such as brand building.
Advertising is often an early use case, since simple things like suppression and lookalike audiences from CDP can generate immediate, measurable result.
It’s usually easy to build a use case for CDP looking at almost any part of the lifecycle, from acquisition through churn reduction.
Confusion about CDP definition can slow down sales, because there are so many different vendors that are hard for buyers to differentiate. The products have changed over time, starting from data capture and now moving to integration with cloud data warehouses and activation in multiple engagement channels, including those outside of marketing.
It’s not clear whether growth in composable and cloud databases has caused a drop-off in selling stand-alone CDPs. CDP Institute industry update report shows that growth has definitely slowed in the past 18 months, but we don’t know the reason. The slowdown might just mirror the general tech employment slowdown post-covid, and it’s possible that the companies buying composable and cloud databases would have built their own solution anyway, so they were never the companies fueling growth of stand-alone CDP.
Composable in APAC is having more of an impact in the small to medium sector than with big enterprises.
AI requires higher data quality.
Companies feel competitive pressure to invest in AI but are moving slowly, in part due to privacy and security concerns. CDP vendors have added both predictive and generative AI to their systems. Predictive AI deployments are much further along and we’re now seeing initial deployments of generative AI.
Generative AI is being used to automate existing tasks but not yet to transform tasks. Expect to see more impact next year in terms of benefiting system users, developers, and ultimately at touchpoints where it will change the customer experience itself.
Generative AI will eliminate some jobs and create others, probably for a net positive impact. People can view AI as a centaur that helps them do their work or a cyborg that takes over their job completely. There are many issues to work out from an enterprise perspective before generative AI becomes widely adopted.

Filling the Gaps for a Composable CDP

2024-05-09T19:24:00.006-04:00

Back in February, I mentioned that CDP Institute had published a Composable CDP Self-Assessment Tool that asks people what gaps they must fill to convert their current systems into the functional equivalent of a CDP. I recently checked how many responses we’ve received, and was disappointed that there were just fourteen. Obviously this isn’t enough to draw statistically meaningful conclusions, especially when you bear in mind that the audience of CDP Institute members can’t be considered representative of the industry as a whole. But I summarized the answers anyway to see what rough patterns might emerge. They were much more intriguing than I expected.

To set the context: the survey asks the status of 102 customer data management functions in the respondent’s current systems. These are grouped into eleven categories: data capture, data sources, ingestion, data preparation, data storage, identity linking, customer profiles, data sharing, process integration, segment creation, and segment (i.e., audience) output. Answer options are: not needed; needed and available; needed and not available; and don’t know. The most important of these is “needed and not available,” since that’s a gap to be filled. All questions were required.

When I took my first overview of the data, the first, and critically important, observation was that there did seem to be significant clusters in the answers. That was important because it suggested that people were giving accurate answers – at least, to the best of their knowledge – rather than randomly filling in a response. Had the answers been random, I would have expected to see roughly the same distribution of responses to each question, and that was definitely not the case. I’ll guess that requiring answers to all 102 questions filtered out the people who didn’t take the survey seriously.

The next observation was that gaps were more common in some areas than others. The gap percentages (i.e., percentage of "needed and not available" replies) ranged from 46% to 29%. The pattern is clear: the areas with fewest gaps (data sources, data store, data capture, and data sharing) are needed for pretty much any data warehouse, while the most gaps were tied to customer data management (data preparation, identity linking, ingestion, segment output, customer profiles). This adds to the credibility of the data, even allowing for some confirmation bias. More important, it's a reminder that you can’t assume a data warehouse built for other purposes will have all the features needed to support CDP applications.

A look individual functions shows something hidden by the category averages: even categories closely related to data warehouses will still have functional gaps for supporting a CDP. For example, while data warehouses are generally good at storing data, they often lack third party data and privacy management. Similarly, although capturing and sharing data are core capabilities for a data warehouse, they often lack real-time connectivity. This reinforces the need to look at requirements in detail when assessing the suitability of your existing warehouse as the base for a CDP.

It's also worth looking at the functions without the category filters. This shows more clearly that common data warehouse functions (such as loading structured data) are rarely gaps while CDP-specific functions (such as connections to advertising media, anonymous-to-known profile conversations, and end-user data access) are often gaps. There are too few responses to make these rankings anything approaching precise. But, even if they were, what any individual company needs would still depend on its own situation.

Given the small sample size of this data set, it’s best to view these results more as anecdotes than an industry survey. But I still believe they offer some useful guidance to CDP vendors and users:

For vendors: it’s worth noting which categories show the most gaps. The “composable CDP” discussion was initially driven by companies offering reverse ETL features, which correspond roughly to the data sharing, segment creation, and segment output categories. Ironically, these fall in the middle of category gap rankings. The biggest gaps, and probably the greatest opportunity for CDP components, relate to specialized data preparation, identity linking, ingestion and segment output. I'd guess that the reason the "composable" discussion started elsewhere is tools for specialized data prep, identity linking, etc. are already widely available, so there was less room for new entrants. By contrast, reverse ETL is a relatively new category where it's easier for a new firm to establish itself. CDP is actually just one application of reverse ETL; in "crossing the chasm" terms, it's a beach head where new vendors can establish themselves and then grow beyond.

From a different perspective, it's also not surprising that more composable CDP vendors are offering products to fill gaps in other CDP categories: it's always easier to sell a new product to an existing customer than to sell an existing product to a new customers. Offering a more complete set of components begins to make the composable products look more like an integrated suite, losing some of the flexibility of swapping components but also reducing the integration burden that composable vendors otherwise ignore. It still maintains the core difference between a standard CDP that builds a separate database and a "composable" one that draws from the warehouse.

For users: the list of common gaps is a helpful reference for ensuring that problematic functions are all considered in a requirements analysis. This is especially important at companies with strong existing data warehouse teams, whose members may not be familiar with CDP requirements. To the extent that users' own systems match the category gap averages, it may make the most sense to consider custom enhancements to existing capabilities for categories with relatively few gaps, while buying new packaged components for categories with many gaps. Of course, if a company has many gaps across all categories, it probably makes more sense to buy a traditional CDP than to buy, integrate, and maintain a large number of separate components.

CDP Success Depends on More than Marketers

2024-05-01T19:20:00.001-04:00

The CDP Institute runs periodic roundtables where our members discuss industry issues. We had one earlier this week on the topic of helping companies make use of an existing CDP. We chose that topic because we frequently hear that many users don’t know what to do with a CDP after it’s built.

The initial discussion followed the more or less expected path towards solutions such as developing change champions and publicizing success stories to encourage CDP use. Participants stressed the need for organizational change to allow applications that the CDP makes possible. This change includes cooperation between teams that had previously not worked together, new processes, and new metrics for cross-channel marketing programs. We heard that traditional direct marketing organizations are especially likely to struggle precisely because they have such mature single-channel techniques in place.

On the bright side, the group said that real-time capabilities were the most likely to create new opportunities that would make users excited enough to adopt new approaches. Good to know.

The discussion then turned in an unexpected direction: the shift in CDP leadership from marketers to IT and data teams. One participant said in the past two years, he had found that 90% of new projects were initiated by IT. While that’s certainly a trend I’ve observed, the group added several new insights.

One reason for the shift is that CDP deployment requires more technical skill than most marketing teams can muster. I have generally argued the reason for greater IT and data team involvement is the broader use of customer data beyond marketing departments. This makes the CDP an enterprise-level project. But it’s probably also true that CDPs used more widely throughout the organization are more technically complicated than CDPs used only in marketing. Broader applications imply more data sources, more types of processing, and connections to more applications. This all requires greater involvement by the company’s IT and data teams.

IT and data teams need new skills to succeed with a CDP. This goes beyond the need to understand specific processes such as identity resolution, which many IT and data teams had not previously needed to support. It includes the need for those teams to better understand how marketers actually do their jobs. Without that fundamental understanding, IT and data teams will make poor choices in designing systems to support marketers. Traditional project-by-project requirements gathering isn't enough.

More specifically, IT and data teams need to recognize that marketers have a different attitude towards data. One participant observed (and I agree) that IT teams want to make sure everything is correct before they release any data, while marketers are willing to accept some errors in return for quicker results. It’s an astute observation about culture that both sides should recognize and discuss so they can agree on the right balance for any particular application.

A bigger role for IT and data implies a smaller role for marketing technologists. They’re still involved in selecting martech, of course, but are less likely to lead a CDP project than they once were. This probably applies to other marketing technology projects as well, especially if it applies to departments outside of marketing. This isn't a death knell for martech staff. But it does imply a shift in responsibilities, away from designing the martech stack to system administration, support, and analytics.

Those final two points – about the different attitudes towards data quality and the changing role for martech staff – both point to a need for more CDP education aimed at IT and data teams, rather than just marketers. That’s something we’ll consider here at CDP Institute and I suspect will be important for industry vendors, consultants, and practitioners as well.

CDP Overview: How We Got Here, Where We're Going, and What Could Get in the Way

2024-04-26T12:23:00.004-04:00

Has anyone asked you recently about the past, present, and future of Customer Data Platforms? No? That's odd; people ask me those questions

all the time. Here are my current answers.

It’s eleven years since the term “customer data platform” appeared in 2013. What stages has it gone through over that time?

The first stage was just to recognize that CDP was a separate category of software. What was new about CDP was that it was packaged software that was building a customer database. Before then, customer databases were custom projects, like a data warehouse, and the only packaged marketing software was applications like campaign management systems or predictive modeling tools. The earliest CDPs actually bundled the database building capability with an application. But, fairly soon, vendors realized that there was more value in the database building features, which were rare, than in the applications themselves, which were fairly common.

Once we had identified the CDP category, the next stage was convincing people that it was important. The concept itself was easy to grasp (“all customer data in one place”). But there was initial skepticism about whether it was a legitimately separate category or just another name for existing technologies such as CRM, data warehouses or DMPs. So we spent most of our time explaining the difference between CDP and other systems that also built customer databases. In fact, the formal CDP Institute definition – "packaged software that builds a persistent, unified customer database accessible to other systems" – is carefully crafted so that each term identifies a differentiator between CDP and some other type of system. “Packaged software”, for example, distinguishes CDP from data warehouses and data lakes, which are custom built. I won’t go through the other elements point-by-point.

Once the concept was reasonably well defined, we faced skepticism about need for separate CDP database, compared to just reading existing source data in real time to build profiles. That’s what ”persistent” refers to in the CDP definition. It took the big martech vendors including Salesforce and Adobe several years to accept that. The reason is that building profiles on demand by assembling data in real time just takes too long.

Even after category was established, there was great confusion about what really qualified as a CDP. That’s why the CDP Institute launched its RealCDP certification program, which expanded on our definition by defining five key requirements: assemble all data types, keep all details, retain the data indefinitely (subject to regulatory constraints), build unified profiles, and share the data with other systems. We later added a sixth point relating to two real time capabilities: access to individual customer profiles, for things like call centers and website personalization, and real time event triggers, for things like responding to dropped shopping carts. Of course, we are just one voice among many, and are easily drowned out by vendor promotions. So, unfortunately, some of that confusion persists to this day.

Part of the reason for the continuing confusion is a debate over whether CDP just builds profiles or also should include data activation capabilities such as analytics and personalization. Our view is they’re not essential, but over time, we’ve seen the industry split between CDPs that only build profiles and CDPs that include activation functions. The majority of CDP vendors provide activation, so that’s clearly what most buyers want.

More recently, we’ve seen interest in using customer data beyond marketing, which makes IT and data teams more interested in CDP projects. That’s a big change because when CDP was just a marketing tool, IT teams were very happy to let marketers buy the CDP becaise it kept the marketers happy without consuming IT resources. Now, CDP is too important to be left to marketers. This also makes the CDPs that just build profiles more appealing in some situations – and we have in fact seen slightly faster growth in that group most recently.

IT involvement in turn brings more interest in companies building their own CDP, and leveraging their existing data lakes and warehouses as the foundation. This has been called ‘composable CDP’ although it’s not really the right use of the term ‘composable’. Most people now agree that 'warehouse-based' is a better label. Semantics aside, the problem is that IT teams often underestimate the requirements for a proper CDP and, thus, the work involved in creating one. So it’s important to ensure they do a thorough job of scoping the project in advance, so they make the best choice.

The other thing we’ve seen, more or less continuously, is expansion of CDP into new industries. Originally they were used primarily in retail and media. Then, they grew more common in financial services, hospitality, and telecom, which are all industries that traditionally had pretty good customer data systems. Most recently, we’re seeing CDPs in education, healthcare, and government applications.

We’re also seeing CDP used more for advertising applications, as companies lean more heavily on first-party data to replace the loss of third-party cookies and replace other targeting methods that become harder as privacy rules become more stringent.

In fact, CDP also supports other privacy-related applications, such as closer control over ensuring that contacts are authorized by consumer consent, and providing data clean rooms for privacy-safe data sharing. Funneling all customer list creation through a CDP is one way to avoid breaking privacy rules.

Where do you see the industry headed next?

The biggest issue right now is the movement towards ‘composability’. We can’t control whether IT and data departments try to build their own CDP-equivalent. But we can educate them about the actual requirements and we can give them tools that make it easier to meet those requirements. Many CDP vendors are now breaking apart their systems to provide modules that companies can use as components in building an internal system. Of course, that helps those vendors to survive a transition to a composable CDP world, but it also helps to maintain the reputation of the industry as a whole. What we don’t want to see is composable CDP projects fail because that makes people question the value of the CDP concept itself.

Closely related to the composable trend is a trend for ‘no copy’ access to external data by a CDP. This means the CDP can read data from other systems without copying it into the CDP data store. That used to be more or less impossible at scale, because finding, reading, and integrating masses of external data took too long. With today’s technologies, that process can be much faster, so it becomes a more practical alternative. Some common use cases are reading things that change quickly and are only relevant at the moment you need them, like inventory levels or local weather conditions. Of course, this ability also blurs the distinction between a traditional CDP, which imported all its data, and a warehouse-based CDP, which works with data stored externally. That’s okay but it does add still more confusion to the discussion.

I’ll also make a side note that the original CDP skeptics argued for reading source system data on demand, so it may seem this proves they were right. But so far I don’t think you can have a CDP that only works by reading external data on demand, because some processes like identity resolution and data aggregation still take too long to do purely on the fly. So I see the future CDP as having a core of data that it does copy and store in its own database, for things like the identity graph that tells how to combine data from different sources into each customer’s profile. Whether that database is a CDP or data warehouse doesn't matter: either way, the data will have been copied from the original source system and preprocessed for use by the CDP. The core data could well be supplemented with data read on-demand from external systems, which could be the original source or a data warehouse or data lake. At least in theory, this would give you the best of both worlds: minimal data copying but maximum performance.

Putting aside composability, CDPs will need to adjust to new privacy rules by adding features like encryption, advanced privacy policy management, and data clean rooms. They’ll adjust to new media and new data types, like audio and video, which need to be not just stored but analyzed in ways that are currently close to impossible. And, of course, they’ll adjust to growing AI capabilities, which will make it easier to perform some functions like adding new data sources, matching identifiers that belong to the same person, building predictive models, and analyzing business results. AI will also add to the volume and complexity of data that CDPs have to handle, which itself may require new technology to support. For example, if AI starts to create a huge variety of content personalized for each individual, that makes result analysis vastly more complicated. Somehow the CDP will need to deal with that.

On a more prosaic level, I expect to see more CDPs that are tailored to the needs of particular industries. That’s a typical development for a mature software category. Specialist systems can be more cost-effective to build and deploy, because they use special data structures and features tailored to a particular industry’s needs. This will bring down the cost of CDPs, which has been an obstacle to expanding the base of CDP purchasers.

And, as I’ve already mentioned, I expect to see CDPs used more widely beyond marketing. One thing we’ve learned in recent years is that customers expect a personalized experience every time they interact with a company, whether it’s before they buy a product or after they start using it. That requires making customer data available at every interaction point. Beyond direct customer interactions, teams like product development and operations planning also can benefit from using customer data.

Are there any particular obstacles or threats to the CDP’s continued success?

Certainly CDPs will have to adapt to the changes we’ve already discussed in data types and volumes, in the users for customer data, and in whatever technical developments change the best way for CDPs to be built. Individual vendors may struggle to keep up. But I think the need for complete, sharable customer profiles is here to stay. So to the extent that that’s the core definition of a CDP, the future of the industry is secure.

That said, I do see two major threats to the CDP industry as we know it today.

The first is technical: the cloud databases from specialists like Snowflake and Databricks and general cloud platforms like Google Cloud, AWS, and Microsoft Azure. All those vendors are increasingly eyeing the giant pile of customer data held in a CDP and wanting it for themselves. To get it, they’re adding applications to build profiles, such as data quality and identity resolution, and to do analytics and marketing. Sometimes they make the applications to be features to their own database systems; sometimes they build direct integrations with specialist applications through a marketplace of some sort. Either way, this cuts out the CDP, which has traditionally been an intermediary between enterprise data stores and business applications. Again, this comes back to the notion of the warehouse as the primary customer data store, which makes the CDP database unnecessary. I think threat is limited today because most IT departments don’t have the resources to build those systems for themselves. But as new tools make it easier to build those systems, the threat becomes more important. There are really just two ways for CDPs to respond to this threat:

one is to become platforms themselves, which is to say, to replace the data warehouse itself. That’s not as crazy as it sounds, since the CDP is really a set of tools that builds the customer database, so it can work on top of a Snowflake or Google BigQuery or whatever database the client wants. In this world, the CDP evolves from a tool to build customer profiles to a tool to build data warehouses in general. Difficult but not impossible, at least for the largest CDP vendors who have the resources to compete.

The other option is to become applications on top of the warehouse, offering specialized capabilities like cross-channel journey orchestration. The would build on CDPs’ existing capabilities to share their profiles with activation systems, and to themselves do activation functions that apply across all channels, such as predictive models and best offer selection. It’s actually an easier path for most CDP vendors, since it continues their role as intermediaries between data sources and business applications. But it’s also a fairly narrow role and there will be lots of competition. Specialization by industry, company size, and other variables will be the key to avoiding that competition by becoming the best choice in a particular niche.

The second threat isn’t technical, but the ability of organizations to actually make use of a CDP once it’s built. We run into this all the time, when companies tell us their staff doesn’t know what to do with a CDP or doesn’t have the skills to do what they’d like. It’s a problem that will only get worse as customer data grows more complicated and there are more possibilities for using customer data. Maybe AI will solve the problem, and that’s something CDP vendors need to invest in to make happen. But there’s no guarantee that AI will be the answer, and my guess is that even AI will need skilled users to take advantage of the possibilities it creates.

Of course, if the CDP turns out not to be useful, the staff members will blame the CDP, not their own lack of skill. But it doesn’t really matter who takes the blame: if CDPs don’t create value, companies won’t be willing to invest in them. So it’s really important for the CDP industry to help train CDP users, not just in the technical details of how to use a CDP, but in the business programs that a CDP can support.

Send in the Clouds: Martech Moves to Cloud Platforms

2024-04-15T22:12:00.003-04:00

Last weekend brought the intriguing rumor that Salesforce is in late stage talks to acquire data integration vendor Informatica. This followed the previous week’s rumor that Google-parent Alphabet is consideringan offer for Salesforce-competitor HubSpot, and came the same week as a slew of partnership announcements tied to Snowflake’s Marketing Data Cloud Forum.

You don’t need a crazy wall to see how these events are connected. Cloud database vendors including Google and Snowflake are expanding into marketing applications. This poses an obvious threat to customer cloud vendors like Salesforce, in good part because it expands the ability of IT and data teams to build their own solutions rather than buying them. Viewed in this light, Informatica helps Salesforce to compete by strengthening its ties with IT and data teams. Of course, this is the same benefit that Salesforce sought with its Tableau, Slack and Mulesoft acquisitions, although you could argue those were all primarily tools for business users while Informatica is used primarily within IT and data teams.

A stronger strategic argument is that an Informatica that is properly integrated with Salesforce Data Cloud would help those teams build more powerful databases within Salesforce. This would make Data Cloud a more viable alternative to Google Cloud or Snowflake as the company’s primary customer data store. In Salesforce’s wildest dreams, Data Cloud might even replace data warehouses.

Let’s put aside the question of whether Salesforce could or should become an enterprise data warehouse supplier. The immediate question is whether it can retain its position as a primary platform for customer data and customer-facing applications, or that role will be taken over by cloud platform vendors like Google Cloud, Amazon Web Services, and Microsoft Azure and cloud database specialists like Snowflake and Databricks.

I don’t know the answer. But what does seem clear is that customer systems will run on platforms, whether from Google, Salesforce, Snowflake, or someone else.

The platform approach isn’t particularly new but it’s not the way things have always been. Remember that most customer systems – even today – are stand-alone, self-contained products that maintain their own databases. These are the dreaded silos we keep hearing about. The trick is to connect these silos, initially at the data level by copying their data into a central warehouse or Customer Data Platform, and ultimately at the decision and delivery levels through shared journey orchestration and messaging.

Platforms solve the data problem, since all systems run on a common data store. They don’t necessarily unify the decision and delivery layers, since those functions are handled by applications built on the platform. Those boundaries are not clear: platform vendors sometimes add decision and delivery functions. Indeed, the risk that the platform vendor will incorporate your application is one that application developers must accept as the price for accessing the platform vendor’s customers. Data clean rooms are a good example: once independent applications, they are now offered as core platform services. Machine learning and artificial intelligence are following the same path.

The transition from silos to platforms is far from complete, but the industry direction is clear. The entry of cloud and cloud data platforms as alternatives to customer clouds is likely to accelerate the change. As we’ve already seen, platforms are a mixed blessing for application developers: they gain access to a larger audience but risk of being undercut by a platform product extension. (The good news is those extensions often involve acquiring a leading application.)

Application vendors also benefit from the platform providing data management functions that the application vendor would otherwise have to build for themselves. This simplifies application development, freeing resources to improve the primary application functions. Unfortunately, it also makes it easier for other companies to build competing applications. This lower barrier to entry (and to continued survival) is one reason the number of marketing applications has increased so sharply in the past decade.

The only way for application developers to escape this "app trap" is to themselves become platforms. Many aspire to this and some, including HubSpot, have had some success moving down that path. But most are too small to support a robust app marketplace and certification program or to attract a critical mass of application vendors.

And what does all this mean for marketers and other business users? Since the platform model isn’t new, we can already draw on experience. It’s a mixed bag: buyers have a greater choice of applications for any particular task, which is mostly good but does create a burden of having to choose. Buyers are tightly bound to whichever platform they select, since switching platforms means switching the related applications as well. This creates a quasi-monopoly situation, with the potential for higher prices and poorer service that comes with it. Raise your hand if you’ve seen that already. This will only become worse as platforms become more common, making it harder for independent application vendors to survive and, thus, harder for companies to assemble their own ‘best of breed’ architecture from separate point solutions. In other words, you may miss those silos when they’re gone.

It doesn’t have to be this way. The alternative to a platform architecture is a true composable architecture, where different applications are written to common standards that are not controlled by any particular platform. The ecommerce world has managed this to some extent, with ‘headless’ solutions and standards enforced by the MACH alliance. Standard data access languages like SQL are another example. That the customer data world will come up with similar standards seems doubtful, although there’s always hope.

It’s also possible that integration platforms will let different systems interoperate even without shared standards. That’s certainly their goal but so far results are limited by the effort to accurately capture all the nuances of one system and present them to another without losing anything in translation. AI might well change things by making this translation easier. But AI might change a lot of things, so let’s not count on any of them quite yet.

For the immediate future, then, we can expect a handful of platform vendors to control customer data at a growing number of companies. Those companies will be captives of the platforms, but it will be a fairly enjoyable captivity, with many pleasing applications to choose from and few obstacles to working as they please. Every so often they’ll pull back a curtain and see the bars on the windows. Some will be so unhappy that they'll escape. But most will accept the limits of their chosen platform and get on with their work.

Composable vs Packaged CDP: How Can We Help?

2024-02-16T08:50:00.052-05:00

“Composable CDP” has been consuming much attention at CDP Institute as we wrestle with how best to help buyers who are considering it as an option. The Institute published a Composable CDP Self-Assessment tool a few weeks ago, which gives a checklist of the functions required to replicate the profile-building capabilities of a standard CDP. This was intended as the centerpiece of our Composable CDP Knowledge Hub but has attracted fewer users than expected. The explanation I find most plausible is that few people can answer the Self-Assessment’s detailed questions about existing systems and requirements.

In itself, this isn’t terrible: much of the value in the Self-Assessment comes from giving users a comprehensive checklist of items to consider, thereby highlighting the true scope of a Composable project. Still, it’s clear the Self-Assessment isn’t helping buyers as much as we had expected, so we’re considering what other tools might be more useful.

The discussions surrounding this have helped to clarify my own understanding of where Composable CDP fits into the greater scheme of things. The key insight is that the choice to use a Composable or packaged CDP is ultimately a tactical decision that addresses just one stage of the CDP project. It doesn’t affect the previous stage, which is requirements definition, or the following stages, which are deployment and activation. To overstate the case just a bit, marketers and other CDP users don’t care how their CDP gets built; they just want to use the resulting profiles to do their jobs. To the extent that CDP users do care about the Composable vs packaged decision, it’s because that choice impacts the cost, timing, and quality of their system.

This insight in turn clarifies the distinction between knowing enough to make the Composable vs packaged decision, and actually deciding which is the best choice. The knowledge needed to make the decision includes:

Defining business requirements, which requires business users who understand their needs.
Understanding existing systems, so you can identify gaps that must be filled to meet business requirements
Understanding the organization’s capabilities for filling the gaps with either a Composable or packaged solution. These capabilities relate to technical resources including staff skills and budgets. I believe that assembling a Composable solution requires more extensive technical resources than buying a packaged CDP, although some Composable advocates might disagree.
Estimating the costs of delivering a Composable solution and a packaged solution so you can compare them.

The Self-Assessment tool addresses only the first two of those items: it asks users what they need and what their current systems can provide. So, even if the users could answer all the questions, it wouldn’t provide all the information needed to make a sound decision. Again, this isn’t exactly a flaw, since those are two important types of information. But it is a limitation.

Only after all the information listed above has been gathered can the company address the second question of whether Composable or packaged is its best choice. Answering this question depends on the combination of capabilities and costs. That is, for Composable to be a viable option, the company needs to have adequate technical resources to assemble a Composable solution, and for Composable to be the best option, it has to offer the best combination of cost and value.

(If we assume that either option delivers the same value, then the only difference is cost. In theory, every system that meets same requirements will deliver the same value. But in the real world, there will be differences in the quality of the resulting systems. These will depend on the specific tools that are chosen, not simply on whether the solution is Composable or packaged. This means that buyers must evaluate specific alternatives for either approach.)

Circling back to the original question of how the Institute can help users who are considering a Composable CDP, it’s clear that few people will have all the information they need available when they start the process. So the best we can do is to provide a framework that identifies the four kinds of information to collect and, perhaps, provides a place to store it over time. That could be as simple as a spreadsheet, which isn’t very exciting from a technical standpoint. But if it’s what our members need the most, then that’s what we’ll deliver.

Addendum: Taking things a step further, here's what the tables described above could look like. It would be interesting to collect separate answers from business users and IT/data teams, who very likely would disagree in many places. I could easily convert these tables into an online format and have the system do some light analysis, including a comparison of answers from business users vs IT/data teams. But we'd still run into the same problem, that it takes time assemble answers. So this pushes me back to a spreadsheet that people can fill out at their leisure. You can download that here.

1 & 2: Requirements & Existing Systems
Data Sources	Not needed	Needed and Fully Available	Needed and Partly Available	Needed and Not Available	Don’t Know
- Website
- CRM
- Data warehouse/ - data lake
- Email
- Third party data
- eCommerce
- Web advertising
- Social media advertising
- Point of Sale

Profile Building Functions
- Capture data
- Ingest data
- Prepare data
- Store data
- Link data
- Build profiles
- Share profiles
- Integrate profiles
- Segment profiles

Activation Functions
- Audience selection
- Predictive analytics
- Campaign definition
- Real time interactions
- Cross-channel orchestration

3. Resources
	Rating (1=low, 5=high)
- Customer data infrastructure
- IT/Data team: customer data experience
- IT/Data team availability for this project
- Business users: customer data experience
- Business user availability for this project
- Training resources
- Measurement resources
- Budget for this project
- Management Support

4. Cost
Hard costs (enter dollar amounts):	Composable	Packaged
- Licenses/Vendor Fees
- Labor/Development & Integration Costs
- Operations (hardware, software, on-going maintenance & upgrades)
Other Considerations: (1=disadvantage, 2=same, 3=advantage)
- Time to Value
- Delivery Risk (time, cost overruns)
- Performance Risk (scalability, performance)
- Roadmap (control over features)
- Security/privacy (control over data)
- Quality (best components, competitive advantage, vendor, usability/consistent interface)
-

Does CDP Need a New Definition?

2023-10-29T18:22:00.002-04:00

The earliest Customer Data Platform systems were introduced before 2010; the term CDP was coined to describe this emerging class in 2013. My definition had changed very little when we launched the CDP Institute in 2016, and has been the same ever since: “packaged software that builds a unified, persistent customer database accessible to other systems”. The Institute added the RealCDP checklist* in 2019 to attach more specifics to the definition in the hopes of helping buyers ensure a system that called itself a CDP could actually support the use cases they expected a CDP to support. By then, industry analysts were beginning to offer their own definitions which, while worded differently, were broadly consistent with the Institute definition. Even the major marketing suite vendors, who initially argued a separate (“persistent”) database wasn’t necessary, eventually discarded that position and introduced products that matched our criteria.

A successful concept like CDP quickly takes on a life of its own. It soon became apparent that many people were using CDP in a much looser sense to mean any system that built and shared customer profiles. This extended past packaged software to include custom-built systems and included systems whose scope was more limited than a true CDP. This expansion skewed some survey resuls but otherwise seemed relatively harmless; in any case, resistance seemed both pedantic and futile. What really mattered was these systems still gave CDP users access to the unified profiles.

Unfortunately, the evolution of the term didn’t stop there. As CDP became popular, many vendors adopted the label whether or not they actually met even the looser definition. At the same time, legitimate CDP vendors offered additional capabilities to analyze and deploy the data in the profiles. The resulting confusion ultimately led some vendors to avoid the CDP label entirely because it no longer provided a useful way to differentiate their products. Today, vendors seeking the latest label are more likely to call themselves digital experience managers than a CDP, even if their products meet the CDP requirements.

But the greatest challenge to the utility of the CDP label arose in the past few years when a number of vendors chose to claim that the core feature of the CDP – a dedicated database built by importing data from other systems, a.k.a. “persistence” – could be abandoned while still calling the result a CDP. Their argument was the customer profiles could reside in a general purpose data warehouse, which most companies already had in place.

The claim gained some plausibility from the fact that modern cloud data warehouse technologies, such as Snowflake, Google Big Query, and AWS Redshift, are in fact used by some conventional CDP vendors. The problem was they implicitly assumed that every company’s data warehouse had organized the data into useable customer profiles. In fact, very few data warehouses perform the specific tasks, most notably identity resolution, customer-level aggregation, and real-time response, needed to support CDP use cases. While it’s technically possible to add those features to an existing data warehouse, it’s usually a major project that often costs more, takes longer, and delivers less useful results than installing a separate, conventional CDP. (As always, the details depend on the situation.)

One positive result of the interest in warehouse-based profiles has been the decision of some CDP vendors to break their systems into modules that let users buy the data preparation functions separately from the rest of the CDP. This lets companies that want a warehouse-based system to still benefit from the mature capabilities those CDP vendors have developed over many years. These vendors have also often added the ability to combine data from a warehouse or other external system with data stored in a conventional, persistent CDP database, without actually loading that external data into the CDP. This gains some advantages of the warehouse-based approach, such as reduced data movement and storage costs, while retaining the benefits of the persistent CDP, such as greater control and flexibility.

You’ll note that all these changes affect the input side of the CDP data flow. It was once a very simple process: data from other systems was copied into the CDP, where it was formatted into profiles and shared with other systems. Now, the input process may be some mix of copying data into the CDP, reading fully-formed profiles from a warehouse, or combining internal and external data. By contrast, the delivery side of the process has remained the same: the CDP shares its profiles with other systems.

In some cases, this has led to a subtle shift in perception of the CDP’s purpose: from a system that builds customer profiles, to a system that delivers those profiles to other systems. In this view, the core function of the CDP is to convert general purpose profiles into the specific formats needed by analytical, orchestration, and delivery systems (which we can call “activation systems” if you’ll trade some jargon for simplicity). This may still involve some data processing, such as advance calculation of model scores and aggregates to make them available in real time, and new data structures to hold the results of that processing. So there’s more happening here than a simple data transfer or API call.

Some people carry this shift even further and argue the CDP should really be defined as an activation system that reads profiles from somewhere else. Since three-quarters of the CDPs we track actually do have at least some activation capabilities, this isn’t quite as crazy as it may sound.

All that said, I’m not yet ready to redefine CDP. “Packaged software” and “persistent customer database” are meaningful terms that distinguish one configuration from other approaches (“custom software” and “external profiles”) that can also make profiles available to other systems. More important, the packaged/persistent configuration has significant cost and execution advantages over the custom/external alternative. So it’s important to avoid blurring the distinction between the two approaches.

At most, I’d offer retronyms that distinguish the “packaged CDP” (packaged/persistent) from the “warehouse CDP” (custom/external). By definition, the “warehouse CDP” doesn’t build its own profiles, so the term seems to ignore the fundamental function of the CDP. You might call that oxymoronic, if not just plain moronic. But if we slightly redefine the core function of the CDP to be delivery of profiles rather than creating them, we can overcome that objection and, perhaps, give the market terms it intuitively understands.

You’ll also note the shift from profile creation to delivery puts the emphasis on the delivery side of the CDP flow, which we’ve noted has been the most stable and applies to both the packaged and warehouse approaches. This lets us join the trend to redefining CDP as a profile sharing system.

In short, the key distinction is whether the primary customer profiles are built and stored in the company data warehouse or in a separate CDP database. It’s worth maintaining that distinction because one approach relies on the company’s technical staff to assemble the functions needed to build and store the profiles, while the other relies on a packaged CDP to provide those functions. There are variations within each theme, including whether the warehouse uses modules from a CDP vendor to assemble its profiles or to help deliver them, whether real-time data is posted to the warehouse or held separately, and whether the CDP enriches its profiles with external data without importing that data. These are important from a practical standpoint, but do not affect the fundamental architecture of the system. Focusing first on where the profiles reside should help buyers understand the most important choice they have to make. This choice will in turn determine the other decisions they have to consider.

_________________________________________________________

* ingest data from all sources, retain all details, keep the data as long as desired, build unified profiles, share the profiles with any other system, and enable real-time event-triggers and profile access.

Do Self-Service Systems Really Lead to Better Results? Our Member Survey Offers Surprising Answers to Industry Questions

2023-09-18T14:33:00.011-04:00

The CDP Institute just published its annual Member Survey, which is always a treasure chest of interesting data. I’ve already published my primary analysis on the Institute site (you can download it here) but wanted to call out a number of findings that either contradict or confirm martech industry conventional wisdom. After all, nothing’s more fun than tweaking the nose of authority.

Data unification is growing: false. Most of us have a deep-rooted confidence in progress, at least when it comes to technology (human nature is another matter). The need for unified customer data has now been so widely understood for so long that we just naturally assume that more companies will have developed it. That has been true since the survey began in 2017 through last year’s survey: both CDP deployment and presence of a unified customer database have increased steadily. But both measures fell in the current survey. It’s hard to imagine companies have actually abandoned their unified databases, but even if growth has only slowed, that would be a big surprise.

Budget pressures are slowing industry growth: true. Some industry vendors report business is booming, but most will admit buyers are taking longer to make decisions. Sure enough, the fraction of vendors who reported growth in CDP investment is down from last year’s survey, both for the past year and the current year.

Budgets may not be the only reason for slower growth, but the budget pressures rose more quickly than any other obstacle (from 20% to 29%). Cooperation, which can also be a symptom of budget pressures, grew the second-fastest (36% to 43%).

Budget-pressed buyers are making smarter decisions: false. I can’t point to anyone who has made this exact claim, but think it’s implicit in reports that companies have more martech than they need and hope to simplify their stacks in the future.

The survey does show that budget pressures are changing martech selection methods: more are selecting on cost (up from 43% to 54% for operating cost and 42% to 51% for initial cost), while selection on feature sophistication and breadth have fallen the most (26% to 15% and 41% to 24%).

Unfortunately, our surveys have consistently found that selection on cost correlates with low satisfaction with martech investments, while selection on features correlates with high satisfaction. So it seems that a short-term focus on cost is likely to cause long-term problems with martech results.

Martech departments are growing: true. The fraction of survey respondents who said they work in a martech department has more than doubled since the last survey, which was the first that listed martech as an option. It’s unlikely that the number of people working in martech has actually doubled in the past year, but it does seem reasonable to believe that some meaningful fraction of employees have been moved from marketing or IT into a dedicated martech department.

IT staff is playing a larger role in martech decisions: true. Despite the growth in respondents who work in martech departments, the current survey showed a small increase in the fraction of respondents who reported that that corporate IT manages their marketing technology (28% to 30%) and sharp declines in the fraction reporting that a martech team was in charge (43% to 32%) or each department runs its own martech (27% to 18%). This may be another cost-saving measure or it may reflect the growing importance attached to customer data (and martech in general) throughout the enterprise.

The bad news is that IT responsibility also correlates with lower martech satisfaction. Bear in mind that survey respondents are themselves mostly martech and marketing people, who are generally happiest when their own team is in charge. IT people probably give a different answer but are a small fraction of the survey respondents.

CDP projects are easy: false. Past surveys have found that about 60% of deployed CDPs are reported as delivering value, while the remaining 40% are struggling. I have always suspected most of the 40% are new projects that will deliver value eventually. The success rate is much higher in the current survey (80%), possibly because the slowdown in deployment has meant there are fewer new projects. But I'd want to see similar results in one or two other surveys before accepting that as a trend.

A separate question, asking vendors about success rates, finds the fraction reporting that say almost all or the majority of projects are successful has grown from 48% to 54%. While this might suggest some improvement in success rates, the more important message is that a bare majority of vendors say most CDP projects succeed. Even when answers from CDP vendors are tabulated separately, just 68% say nearly all or a majority of projects are successful. Figures are lower for service providers (45%) and other respondents (15%).

It's important to realize this is no worse than success rates for other large system deployments. In fact, it's apparently better than average, since most studies put failure rates at 60% to 70%. (See this page for a compendium.) But these findings should dispel any notion that CDP deployments are easy.

CDP projects fail because of technical complexity: false. It's also important to recognize that CDP failure rates are not due to any inherent problem with CDP technology. As in previous surveys, by far the top reason for CDP project failure is organization. This has grown even more prominent in the current survey, while problems with poor requirements and CDP performance have fallen.

Comparing CDP status with satisfaction offers additional insight. The satisfaction measure reflects success with martech in general, not with CDP in particular. In other words, companies with a high score are "good at martech". So, while it's not surprising that satisfaction rates are high among companies with a successfully deployed CDP and low among those who have not, this does support the position that CDP success is based more on the skills of the organization than CDP technology in particular.

Privacy is growing more important: true. Last year’s survey showed a distressing decline in the priority given to data privacy regulations, with the share of firms making little effort to comply growing from 12% to 20%. That trend has now reversed, with the share of companies using privacy as a selling point increasing from 21% to 27%.

Privacy was also listed in the CDP benefits question for the first time. It was cited by 22% of respondents, ranking seventh of eleven items, and shows a below-average satisfaction score. This may indicate that most CDP users are looking elsewhere for their primary privacy management solution.

Self-service leads to success: false. The martech management section of this year’s survey added a new question about whether companies seek systems that empower business users to execute tasks without technical assistance. The concept of “no-code” is directly relevant here, although we didn't use the term. This capability was by far the most common management technique, which wasn’t surprising: “empowering end-users” is a popular goal that saves money and makes users more effective. But it also correlated with a low satisfaction score, which was surprising indeed. Is it possible that self-service doesn’t save money or make users more effective after all?

Answers to another survey question shed a bit more light. Among the options listed for CDP capabilities were self-service data extracts and self-service predictive models. Self service extracts were a common requirement (41%) correlated with a roughly average satisfaction rating, while self-service models were an uncommon requirement (8%) correlated with a very low satisfaction rating. My interpretation is that self-service extracts are a simple task that’s well understood by business users, so companies with well-run martech operations provide that capability. By contrast, self-service predictive models are a complicated task that few users are equipped to handle on their own, so they are prioritized largely by companies with a poor understanding of what makes for martech success. The larger message is that self-service should be deployed only when users are ready for it – and pushing beyond those limits can cause problems despite the apparent savings in cost and time.

Real-time processing is a high priority for most users: mixed. Real-time processing is commonly cited as an important CDP capability. In fact, the CDP Institute’s RealCDP requirements include real-time access to profiles and real-time event triggers. Again looking at the capabilities question, we see that real-time profiles are indeed a common requirement (42%), while real-time recommendations are much less common (15%) but correlate with very high satisfaction. The lesson here is that there are different types of real-time processing, and users consider some more important than others. Discussions about real-time should include similar nuance.

CDP must load data from all sources and retain full details: mixed. Both of these items are on the RealCDP requirements list. But loading data from all sources is the highest ranked capability (78%) and correlates with roughly average satisfaction, while loading full detail is cited by just 14% of respondents and correlates with much lower satisfaction. As with real-time and self-service, I take these answers to show that users have a mature understanding of what they do and don’t need from the CDP. Loading all data sources is a core CDP promise and goes back to the problem CDP was originally designed to solve: getting access to all customer data. Storing full detail is not needed in most situations. In practice, users choose which data elements are worth the cost. That said, I still believe that users want the option to store any particular details they need.

CDP users want to access their data warehouse directly: false. This year’s survey added a capabilities question about reading data from an external source without loading it into the CDP. This is a hot topic in the industry, both as a way of supplementing a traditional CDP (which maintains its own data store) and as a way of building a CDP-equivalent system that relies only on data in the enterprise data warehouse. Only a small fraction of respondents (14%) were interested in this and that group reported exceptionally low satisfaction levels. I interpret this to mean that, despite all the marketplace noise, there is actually very little interest among knowledgeable users in building a CDP that relies primarily on external data.

Summary

This report contains more bad news about industry growth and budget pressures than I like to see, but the tech sector’s troubles are well known and it’s not surprising that CDP projects should suffer with everyone else. I’m especially reluctant to highlight the data about success rates, because I know it will be taken out of context by vendors eager to promote alternatives to CDPs. I should stress again that the main obstacles to CDP success are organizational, not technical, and that success rates reported here are consistent with tech project success rates in general. Bottom line: CDPs are major projects that don’t always go smoothly but are not more failure-prone than any other major IT effort.

What worries me more than bad news is the hints that companies are making poor decisions. Prioritizing cost over requirements will surely lead to more companies purchasing unsuitable products. Putting IT in charge of martech will almost surely lead to unhappy martech users. While budget issues are unavoidable, poor management decisions are unforced errors. Your company should work hard to avoid them.

All that said, the most interesting results are the ones that challenge conventional wisdom. Do self-service systems really lead to bad results? Is the importance of real-time overstated? Is there really so little interest in reading data directly from a warehouse? These are headline topics in martech today, and are often accepted as truth with little discussion. The questions are worth asking because the reality is almost always more complicated than the simple answers. When is self-service useful and when is it over-used? What kinds of real-time processes are really important? How is external data best integrated with a conventional CDP? Answering those questions will help companies make better decisions and ultimately lead to greater martech success.

Unleashing the Power of Customer Data Platforms (CDPs) and AI: A Game-Changer for Modern Marketing

2023-09-14T01:22:00.000-04:00

For some unknown reason, my last three presentations all started as headlines (two created by someone else) which I then then wrote a speech to match. This isn’t my usual way of working. It does add a little suspense to the writing process – can I develop an argument to match the title? It also dawns on me that this is the way generative AI works: start with a prompt and create a supporting text. That’s an unsettling thought: are humans now imitating AI instead of the other way around? Or have I already been replaced by an AI but just don’t realize it? How would I even know?

The latest in this series of prompt-generated presentations started when I noticed that the title shown in a conference agenda didn’t match the speech I had prepared. When I pointed this out, the conference organizers said I could give any speech I want, but the problem was, I really liked their title: “Unleashing the Power of Customer Data Platforms (CDPs) and AI: A Game-Changer for Modern Marketing”.

The idea of “unleashing” a CDP to run wild and exercise all its powers, is not something I get to talk about very often, since most of my presentations are about practical topics like defining a CDP, selecting one, or deploying one. And I love how “and AI” is casually tucked into the title like an afterthought: “there’s this little thing called AI, perhaps you’ve heard of it?”

And how about “game changer for modern marketing”? That’s an amazing promise to make: not just will you learn the true meaning of modern marketing, but you'll find how to change its very nature so it’s a game you can win. Who wouldn’t want to learn that?

This was definitely a speech I wanted to hear. The only problem was, the only way for that to happen was for me to write it. So I did. Here’s a lightly modified version.

Let’s start with the goal: winning the modern marketing game. The object of the game is quite simple: deliver the optimal experience for each customer across all interactions. And, when I say all interactions, I mean all interactions: not just marketing interactions, but every interaction from initial advertising impressions through product purchase, use, service, and disposal. I also mean every touch point, from the Internet and email through call centers, repair services, and the product itself.

That’s a broad definition, and you will immediately see the first challenge: all the departments outside of marketing may not want to let marketing take charge of their customer interactions. Nor is your company’s senior management necessarily interested in giving marketing so much authority. So marketing’s role in many interactions may be more of an advisor. The best you can hope for is that marketing is given a seat at the table when those departments set their policies and set up their systems. This requires a cooperative rather than a controlling attitude among marketers.

The second challenge to winning at marketing is the fragmented nature of data and systems. Most marketing departments have a dozen or more systems with customer data; at global organizations, the number can reach into the hundreds. Expanding the scope to include non-marketing systems that interact with customers adds still more sources such as contact centers and support websites. Again, marketing will rarely control these. At best, they may have an option to insert marketing recommendations directly into the customer experience, such as suggesting next best actions to call center agents.

The third challenge is optimization itself. It’s not always clear what action will result in the best long-term results. A proper answer requires capturing data on how customers are treated and how they later behaved as a result. Some of this will come from non-marketing systems, such as call center records of actions taken and accounting system records of purchases. Again, those systems’ owners may not be eager to share their data, although it’s harder for them to argue against sharing historical information than against sharing control over actual customer interactions.

But the challenge of optimization extends beyond data access. Really understanding the drivers of customer behavior requires deep analysis and no small amount of human insight. Some questions can be answered through formal experiments with test and control groups. But the most important questions often can’t be defined in such narrow terms. Even defining the options to test, such as new offers or marketing messages, takes creative thought beyond what analysis alone can reveal.

And even if you could find the optimal treatment in each situation, the playing field itself keeps shifting. New products, offer structures, and channels change what treatments are available. New systems change the data that be captured for analysis. New tools change the costs of actions such as creating customer-specific content. These all change the optimization equation: actions that once required expensive human labor can now be done cheaply with automation; fluctuating product costs and prices change the value of different actions; evolving customer attitudes towards privacy and service change the appeal of different offers. The optimal customer experience is a moving target, if not an entirely mythical one.

None of this is news, or really even new: marketing has always been hard. The question is how “unleashing” the power of CDPs and AI makes a difference.

Let’s start with a framework. If you think of modern marketing as a game, then the players have three types of equipment: data systems to collect and organize customer information; decision systems to select customer experiences; and delivery systems to execute those experiences. It’s quite clear that the CDP maps into the data layer and AI maps into the decision layer. This raises the question of what maps into the delivery layer. We’ll return to that later.

First, we have to answer the question: Why would CDP and AI be game changers? To understand that, you have to imagine, or remember, life before CDP and AI. Probably the best word for that is chaos. There are dozens – often hundreds -- of data sources on the data layer, and dozens more systems making choices on the decision layer. The reason there are so many decision systems is that each channel usually makes its own choices. Even when channels share centralized decision systems, those are often specialized services that deal with one problem, whether it’s product recommendations or churn predictions or audience segmentation.

CDP and AI promise to end the chaos by consolidating all those systems into one customer data source and one decision engine. Each of these would be a huge improvement by itself:

the CDP makes complete, consistent customer data available to all decision systems, and
AI enables a single decision engine to coordinate and optimize custom experiences for each individual.

Yes, we’ve had journey orchestration engines and experience orchestration engines available for quite some time, but those work at the segment level. What’s unique about AI is not simply that it can power unified decisions, but that each decision can each be tailored to the unique individual.

But we’re talking about more than the advantages of CDP and AI by themselves. We’re talking about the combination, and what makes that a game-changer. The answer is you get a huge leap in what’s possible when that personalized AI decisioning is connected to a unified source of all customer information.

Remember, AI is only as good as the data you feed it. You won’t get anything near the full value of AI if it’s struggling with partial information, or if different bits of information are made available to different AI functions. Connecting the AI to the CDP solves that problem: once any new bit of information is loaded into the CDP, it’s immediately available to every AI service. This means the AI is always working with complete and up-to-date data, and it’s vastly easier to add new sources of customer data because they only have to be integrated once, into the CDP, to become available everywhere.

To put it in more concrete terms, one unified decision system can coordinate and optimize individual-specific customer experiences across all touch points based on one unified set of customer data.

That is indeed a game-changer for modern marketing, and it only reaches its full potential if you “unleash” the CDP to consolidate all your customer data, and “unleash” the AI to make all of your experience decisions.

That’s a lot of unleashing. I hope you find it exciting but you should also be a little bit scared. The question to ask is: How can I take advantage of this ‘game changing potential’ without risking everything on technology that is relatively new and, in the case of AI, largely unproven. Here’s what I would suggest:

Let’s start with CDP. So far, I’ve been using the term without defining it. I hope you know that a CDP creates unified customer profiles that can be shared with any other system. No need to get into the technical details here. What’s important is that the CDP collects data from all your source systems, combines it to build complete customer profiles, and makes them available for any and every purpose.

Some people reading this will already have a CDP in place but most probably do not. So my first bit of advice is: Get one. It’s not such easy advice to follow: a CDP is a big project and there are dozens of vendors to choose from, so you have to work carefully to find a system that fits your needs and then you have to convince the rest of your organization that it’s worth the investment. I won't go into how to do that right now. But probably the most important pro tips are: base your selection on requirements that are directly tied to your business needs, and ensure you keep all stakeholders engaged throughout the entire selection process. If you do those two things, you can be pretty sure you’ll buy a CDP that’s useful and actually used.

That said, there are some specific requirements you’ll want to consider that tie directly to ensuring your CDP will support your AI system. One is make sure you buy a system that can handle all data types, retain all details, and handle any data volume. This does NOT mean you should load in every scrap of customer-related data that you can find. That would be a huge waste. You’ll want to start with a core set of customer data that you clearly need, which will be data that your decision and delivery systems are already working with. This lets the CDP become their primary data source. Beyond that, load additional data as the demand arises and when you’re confident the value created is worth the added cost.

The second requirement is to ensure your CDP can support real time access to its data. That’s another complicated topic because there are different kinds of real time processes. What you want as a minimum is the ability to read a single customer profile in real time, for example to support a personalization request. And you want your CDP to be able to respond in real time to events such as a dropped shopping cart. Your AI system will need both of those capabilities to make the best customer experience decisions. What’s not included in that list is updating the customer profiles in real time as new data arrives, or rebuilding the identity graph that connects data from different sources to the same customer. Some CDPs can do those things in real time but most cannot. Only some applications really need them.

The third requirement relates to identity management. The CDP needs to know which identifiers, such as email, telephone, device ID, and postal address, refer to the same customer. At CDP Institute, we don’t feel the CDP itself needs to find the matches between those identifiers. That’s because there’s lots of good outside software to do that. We do feel the CDP needs to be able to maintain a current list of matches, or identity graph, as matches are added or removed over time. That’s what lets the CDP combine data from different sources into unified profiles.

My second cluster of advice relates to AI. I’d be surprised if anyone reading this hasn’t at least tested ChatGPT, Bing Chat Search, or something similar. At the same time, there’s quite a bit of research showing that relatively few companies have moved beyond the testing stage to put the advanced AI tools into production.

And that’s really okay, so my first piece of advice is: don’t be hasty. You should certainly be testing and probably deploying some initial applications, but don’t feel you must plow full speed ahead or you’ll fall behind. Most of your competitors are moving slowly as well.

That said, you do need to train your people in using AI. They don’t necessarily need to become expert prompt writers, since the systems will keep getting smarter so specific prompting skills will become obsolete quickly. But they do need to build a basic understanding of what the systems can and can’t do and what it’s like to work with them. That will change less quickly than things like a user interface. The more familiar your associates become with AI, the less likely they are to ask it to do something it doesn’t do well or that creates problems for your company.

Third, pay close attention to AI’s ability to ingest and use your company’s own data. Remember the game-changing marketing application for AI is to create an optimal experience for each individual. The means it must be able to access that individual’s data. This is an ability that was barely available in tools like ChatGPT six months ago, but has now become increasingly common. Still, you can bet there will be huge differences in how different products handle this sort of data loading. Some will be designed with customer experience optimization in mind and some won’t. So be sure to look closely at the capabilities of the systems you consider.

Fourth, and closely related, the industry faces a huge backlog of unresolved issues relating to privacy, intellectual property ownership, security, bias, and accuracy. Again, these are all evolving with phenomenal speed, so it’s hard for anyone to keep up – even including the AI specialists themselves. Unless that’s your full time job, I suggest that you keep a general eye on those developments so you’re aware of issues that might come with any particular application you’re considering. Then, when you do begin to explore an application, you’ll know to bring in the experts to learn the current state of play.

Similarly, keep an eye on the new capabilities that AI systems are adding. This is also evolving at incredible speed. Some of those capabilities may change your opinion of what the systems can do well or poorly. Some will be directly relevant to your needs and may form the basis for powerful new applications. We’re still far away from the “one AI to rule them all” that will be the ultimate game-changer for marketing. But it’s coming, so be on the alert.

This brings us back to the third level of marketing technology: delivery. Will there be yet one more game-changer, a unified delivery system that offers the same simplification advantages as unified data and decision layers? The giant suite vendors like Salesforce and Adobe certainly hope so, as do the unified messaging platforms like Braze and Twilio. The fact that we can list those vendors offers something of an answer: some companies think a unified delivery layer is possible and would argue they already provide one. I’m not so confident because new channels keep popping up and it’s nearly impossible for any one vendor to support them all immediately.

What seems more likely is a hybrid approach where most companies have a core delivery platform that handles basic channels like email, websites, and advertising and supports third-party plug-ins to add channels or features they do not. These platforms are already common, so this is less a prediction of the future than an observation of the present, which seems likely to continue. The core delivery platform offers a single connection to the decision layer run by the AI. This gives the primary benefits gained from a single connection between layers, although I wouldn’t call it a game-changer only because it already exists.

My recommendation here is to adapt the delivery platform approach, seeking a platform that is as open as possible to plug-ins so it can coordinate experiences across as many channels as possible. In this view, the delivery platform is really an orchestration system. Which channels are actually delivered in the delivery platform and which are delivered by third-party plug-ins is relatively unimportant from an architectural point of view. Of course, vendors, marketers, and tech staff will all care a great deal about which tools your company chooses.

While we’re discussing architectural options, I should also mention that the big suites would argue that data, decision, and delivery layers should all be part of one unified product, reducing integration efforts to a minimum. That may well be appealing, but remember that most of the integrated suites were cobbled together from separate systems that were purchased by the vendors over time. Connecting all the bits can be almost as much work as connecting products from different vendors.

And, of course, relying on a single vendor for everything means accepting all parts of their suite – some of which may not be as good as tools you could purchase elsewhere. The good news is most suite vendors have connectors that enable users to use external systems instead of their own components for important functions. As always, you have to look in detail at the actual capabilities of each system before judging how well it can meet your needs.

So, where does this leave us?

We’ve seen that the object of the modern marketing game is to deliver the optimal experience for each customer. And we’ve seen that challenges to winning that game include organizational conflicts, fragmented data, and fragmented decision and delivery systems.

We’ve also seen that the combination of customer data platforms and AI systems can indeed be game-changing, because CDPs create the unified data that AI systems need to deliver experiences that are optimized with a previously-impossible level of accuracy.

CDPs and AI won’t fix everything. Organization conflicts will still remain, since other departments can’t be expected to turn over all responsibility to marketing. And fragmentation will probably remain a feature of the delivery layer, simply because new opportunities appear too quickly for a single delivery system to provide everything.

In short, the game may change but it never ends. Build strong systems that can meet current requirements and can adapt easily to new ones. But never forget that change is inherently unpredictable, so even the most carefully crafted systems may prove inadequate or unexpectedly become obsolete. Adapting to those situations is where you’ll benefit from investment in the most important resource of all: people who will work to solve whatever problems come up in the interest of your company and its customers.

And remember that no matter how much the game changes, the goal is always the same: the best experience for each customer. Make sure everything you do is has that goal in mind, and success will follow.

In a World Run by AI, The Best Data Wins

2023-07-01T17:17:00.002-04:00

Like everyone else in martech land, I’ve been pondering the future of marketing in a world populated with AI. Most research I’ve seen agrees with this Hubspot report that marketers’ top application for generative AI has been content creation (48%), followed closely by data analysis (45%) and learning how to things (45%). So it’s fairly clear that the immediate impact of AI will be to let marketers create vastly more copy, including real-time messages tailored to specific individuals.

Extending this a bit, we can expect those real-time messages to be delivered within ever-more-finely tailored campaign flows (something else that AI can easily generate) or without formal campaign structures at all. These messages will be orchestrated and optimized across all channels, fulfilling the omnichannel vision that has hovered like a mirage on the industry horizon for decades.

(Whether this results in more or fewer martech applications is a separate question. I tend to agree with this GP Bullhound study, which argues that “Any application introduced as an AI-enhanced alternative to an existing application or as a feature to a platform will likely become redundant when that incumbent platform implements the same AI features.” This suggests that fewer specialist martech products will be needed, since truly new applications are relatively rare. Moreover, the productivity benefits of integrated suites are magnified when AI can easily orchestrate tasks within the suite, but not those on the outside. And, AI systems are inherently more adaptable than traditional software, which must be explicitly programmed for each new task. So extending existing applications becomes easier when AI enters the picture.)

Of course, what the humans reading this piece really care about is the role that they will play in this AI-managed universe. (Come to think of it, the AIs reading this may also care more than we know.) We’re frighteningly close to living the old joke about the factory of the future, where the one human employee's job is to feed the dog, and the dog's job is to keep the human away from the equipment.

The conventional answer to that question is that humans will still be needed to provide the creative and emotional insight that AIs cannot deliver. (That’s what ChatGPT told me when I asked.) Frankly, I don’t buy it: just as people who can fake sincerity, have it made, the AIs will quickly learn to mimic creativity and find which emotion-based messages work best.

Still, let’s be a bit optimistic and assume the marketing department of the future includes one human whose job is to feed the AI. Let’s even assume that human adds some creative and emotional value. The net result across the marketing industry as a whole will be that every company produces a wonderfully high and consistent level of marketing outputs. While that’s great in many ways, it also means that better marketing will no longer be a competitive differentiator. It’s a repeat of the situation in manufacturing since the 1970’s, when quality best practices were applied everywhere, so the differences between the best and worst products were often too small to matter. The winners in that world were companies who could use marketing to differentiate what in fact were commodity products. But when AI lets every company produce great marketing, then marketing itself is also a commodity.

So, how will companies compete in this new world? I’ve already argued it won’t be through emotional insight or creativity. I briefly thought that people might do better than machines at adapting to rapid change. The theory was that AI can only be trained on historical data so it will flounder when faced with unexpected events – which are increasingly common. But, let’s face it: humans also flounder in the face the unexpected. It’s not at all clear they’ll be better than machines at predicting abrupt change, recognizing changed circumstances, collecting new evidence, and finding the new best actions. In fact, humans are heavily biased in favor of making decisions based on past experience, so I’d probably bet on the AI.

But all is not lost. I still see two ways for companies, and the humans who run them (for now), to distinguish themselves.

The first is customer experience. If you consider the true commodity industries, such as telecommunications, air travel, hotels, and financial services, what makes customers loyal to one or another provider is rarely the actual product or price. Rather, it’s the way they are treated. As a baseline, customers expect reasonable service, delivered pleasantly. But loyalty is really won or lost when there’s a problem or special request. This gives the company a chance to distinguish itself, either against an actual competitor or against a customer’s expectations of how they should be treated.

Company policies and systems play a large role in what’s possible, but ultimately it’s the front-line employee whose training, attitudes, and choices make or break the experience. In the future, AI will surely play a larger role in managing these interactions. But, as with marketing, this will yield largely similar results because companies will be using similar AI systems. The differentiator will be the company’s people.

That's good to know, but customer experience is rarely under marketers’ direct control. This leaves one final straw for them to lean on: the data used to train their AIs.

Remember, AI programs themselves will be widely available. As with any other technology, the difference in results will depend on how they’re used, not any difference in the technology itself. Once AI systems are fully deployed, most decisions about things like content and program design will be made by the system, limiting the impact of user choice on the outcomes. But the one thing that will remain under users’ control is the data fed into the AI systems. It’s differences in that data that will drive differences in outcomes. In short: whoever has the best data, wins.

This is an area where marketers have a major role to play. They may not control the various internal and external systems who provide customer data. But they will have a large say in what those systems feed into the primary customer data store which, in turn, will feed the AI. Marketers who select the best data feeds will have a more effective AI and, thus, better final results. This will be true even though AI makes data collection easier: that will reduce the technical barriers to data gathering at all companies, but most barriers are actually organizational and budgetary. Those are company-specific decisions whose outcomes marketers can affect.

This may not be the cheeriest news you’ve heard today. Few marketers chose their career because they wanted to fight political battles with IT and customer success teams. But it does mean that many marketers can continue to play a major role in the success of their organizations, even as most of the traditional marketing tasks are taken over by AI. No doubt, the marketers who remain employed will sneak a few of their creative and emotional insights into AI prompts. Maybe their inputs will even make a positive difference. But what will really matter is how good a job they do at feeding the AI with the best possible data. That's what will empower it to deliver better results than the competition.

Remember: in a world run by AI, the best data wins.

Customer Experience Matrix

What Comes After the Composable CDP?

Fitting Agents into the Sales and Marketing Mix

Let's Debate CDP Functions, Not Definitions

Why Agentic Campaigns Are Not The Future of Marketing

What would an AI-native marketing process look like? ChatGPT's answer is better than I expected

How Will GenAI Change the Consulting Industry?

Taking Advantage of AI Search

CDP Relay42 Bought by Supermetrics: The Embedded CDP Trend Continues

Mapping AI-Based Martech Using the STIB Innovation Diffusion Model

The Future of Customer Management

Mapping against the STIB Framework

Application

Content Creation: This is a popular task to automate. It’s a single task but more than substitution because most vendors connect their content generator to response data and use this to automatically optimize content over time.

Infrastructure

Media Integration: this is the output infrastructure that deliver messages created by the hyper-personalization. The tools are steadily encompassing more channels.

Business Models

What We’ve Learned

Implications

Summary

Predicting the Future of AI-Based Customer Management: Lessons from Machine Tool Motors

State of Martech Report: Customer Data Platforms Are Evolving, Not Dying

Rokt Buys mParticle

CDP Lytics Bought By DXP Customerstack

Uniphore Buys ActionIQ: What We Can Learn

Composable CDP is Dead

CDP Round Table: Forget Composable, Are Cloud Databases a Threat to CDPs?

Filling the Gaps for a Composable CDP

CDP Success Depends on More than Marketers

CDP Overview: How We Got Here, Where We're Going, and What Could Get in the Way

Send in the Clouds: Martech Moves to Cloud Platforms

Composable vs Packaged CDP: How Can We Help?

Does CDP Need a New Definition?

Do Self-Service Systems Really Lead to Better Results? Our Member Survey Offers Surprising Answers to Industry Questions

Unleashing the Power of Customer Data Platforms (CDPs) and AI: A Game-Changer for Modern Marketing

In a World Run by AI, The Best Data Wins