Longbow Direct Marketing -

So what will it take to convince you?

Mark Klein — Sun, 19 Apr 2009 11:55:36 -0400

All of us could use some marketing help in today’s tough environment, and there are plenty of marketing services firms ready to sell us their services. But more than ever companies need to spend their money wisely and justify, in advance if possible, commitments they are about to undertake. At the risk of making our own life more difficult, we offer here the minimum set of requirements you should require from any analytics vendor you’re thinking of using.

Marketing insight. Don’t let a vendor try to impress you with tables of data. Make sure they can deliver fresh insights into both your customer population and how you market to them. If the insights surprise you, all the better.
Actionable recommendations. Make sure what the analytics vendor tells you will translate into campaigns you can execute. Be wary if you are being asked to subdivide your customer population into twenty-five segments, from the one or two you had before. Make sure you have the resources to carry out the recommendations.
Proof that you will increase revenues by following their plan. Ask for case studies and back tests before you commit. A reputable vendor will have an excellent idea of what they can do for you after a brief look at your transaction data.

Seems simple enough, doesn’t it? It’s not as easy as it looks, which is why you need to be demanding and persistent. Even with us.:-)

Beware the Cellophane Curtain

Mark Klein — Tue, 03 Mar 2009 11:44:23 -0500

On the shelf next to the fax machine in my office sits a boxed piece of software called GOLDMINE, “The #1-Rated Contact Manager”. Copy on the box says it is “100% Internet Ready!” and “Designed for Microsoft Windows 95”, which gives you some idea of how long it’s been sitting there with its cellophane shrink-wrap still unbroken.

As someone who spent years sweating the advertising that was supposed to make a sale, this makes me heartsick. That’s because, I learned an important lesson along the way. It makes no difference how the application is delivered – it can be downloaded, shipped in a box, or delivered as a web service - it isn’t really sold until it’s unwrapped, installed, integrated into the buyer’s business process, and used regularly.

That is why we’ve just developed a Concierge Service to support our Longbow clients.

None of this has been an issue for the previous nine years at Loyalty Builders, because, until Longbow came along, our services were customized, delivered 1:1, and paid for accordingly. Longbow, however, puts sophisticated analytics in the hands of marketers who’ll customize it for their own use and run it themselves. There’s one little problem though, and it’s not in the application.

We’ve discovered that the biggest barrier to getting started in analytics isn’t learning how to use the application. It’s in pulling together and preparing the data for analysis. Research shows that industry wide, this is by far the most time consuming task (see the chart below). So, going forward, each of our Longbow clients will be assigned a concierge whose job will be to navigate any rough patches (notably, finding and preparing data for upload and analysis) and getting the client up and running as quickly as possible, so that the benefits of analytics begin flowing.

When a vendor gets caught behind the cellophane curtain, he or she has only made a fraction of the sale. There won’t be any sales of accessories. There won’t be any revenue from support or service contracts. There won’t be any terrific word-of-mouth, the most powerful, least expensive kind of advertising. There won’t be any upgrade revenue. Something like Concierge Service has to be offered.

New products change business processes. All sorts of hiccups can happen along the way. But if you can’t get rid of the hiccups you haven‘t really sold anything.

Time spent on tasks

When Rexer Research asked respondents how they allocated time devoted to analytics, they confirmed our view: more time is devoted to accessing and preparing data (green area) than any other task.

Marketing to your best customers when business stinks

Mark Klein — Sun, 01 Feb 2009 13:46:16 -0500

It’s the oldest cliché in business - recession is a time of opportunity. But a raft of research confirms it's true. People who market aggressively in downturns not only survive but consistently emerge stronger. As Casey Stengel would say if he were alive today, “you could Google it”.

The most aggressive approach to bad times is to figure out who your best customers are and market to them relentlessly. Analyze your customers based on transaction data that’s sitting in sales or accounting databases. It’s going to tell you things about your best customers you never realized.

But beware. Best customers are not just the 5% or 10% who generate the most revenue. In fact, customers who spend the most may already be spending all they can with you.

The wonderful thing about segmentation is that when you look at it carefully you’ll realize that you have at least three kinds of best customers:

Best revenue producers
Best growth prospects
Best worst customers (nope that isn’t a misprint)

These are the customers who will help you through this tough market.

Each of these segments behaves differently, as you can see in the segmentation chart. Your campaigns will be more successful when you market to each segment based on what analysis predicts they’re likely to do next. Sending the same offer to everyone is more than a waste - it can damage your brand among people who see your message as a bother not a help.

Best revenue producers - Let them know how valuable they are. Ply them with offers that will help you retain their business and sustain their purchase levels.

Best growth prospects - Mid-tier customers who often spread their business among suppliers. Market to them to win a bigger share of their trade. If they’re buying a narrower list of items compared with others, your analysis will tell you so.   Work to broaden their market basket by identifying products they’re probably buying elsewhere and making offers to win that business.

Best worst customers - New customers who’ve bought once or twice from you don’t rank high, because they haven’t yet generated a lot of revenue.   But some of them are destined to be bigger revenue producers if you treat them right. Identify one-time buyers who’ve bought recently and go after them. A second sale is a big step in securing future loyalty.

Forget about your old definition of best customers in 2009, and I believe you’ll be in better shape by year end.

How to measure marketing effectiveness

Mark Klein — Tue, 13 Jan 2009 13:48:31 -0500

“You’ve looked at our data. So how are we doing? How do we compare with your other clients?”

These were questions thrown at me by one of our newer clients, and they weren’t new questions. Marketers are hungry for an outside, impartial opinion of their efforts. It’s a fair question and a challenge we couldn’t resist.
Whether it’s a long time client or a new prospect looking at their free customer report card, it’s clear to them that we know a lot about their marketing efforts and could offer an informed response.

We decided to accept the challenge and developed a Marketing Effectiveness Score (MES) based on some closely held beliefs and time-honored marketing principles. There are six components to our Marketing Effectiveness Score:

Churn-- the relationship between new customers coming in and existing customers leaving.
Revenue mix-- the ratio of revenue from new customers to existing customers
Revenue growth-- the ratio of the change in revenue compared to the total revenue in a prior period
Revenue at risk--the ratio of the probability-adjusted revenue at risk to the total revenue, a measure of the exposure of the company to customer defection
Share of wallet—a measure of the extent to which customers purchase all of your products and services
Customer loyalty—a profile of the loyalty scores of all customers, and how those scores are changing

These components are aggregated into a number from 0 to 100 that measures your company at one moment in time. Business is a race with a finish line that's always moving. It's not only natural to want to know how you're doing; it’s essential if you don't want to be passed in the race without knowing it. We calculate the MES regularly for our clients so they can evaluate their progress, and once for each company that requests the free customer report card.

So now we can flip that question right back at you: How is your company doing? Do you know your Marketing Effectiveness Score?

When the going gets tough

Mark Klein — Fri, 12 Dec 2008 10:43:12 -0500

We all know these are tough economic times. That message was reinforced for me again yesterday when I got a call out of the blue from a marketing analyst at a company not on my radar screen. She was charged by her VP with setting marketing priorities for next year. Specifically, she was asked to make marketing recommendations to help her company survive during this recession. The analyst was casting a broad net and found me.

After a short discussion, I found that her company’s revenues were excessively skewed more to new customer revenue compared to revenue from existing customers, so a sure answer to her question is to ramp up existing customer marketing. The justification for that effort is clear: you know who the customers are, you can figure out what they are likely to buy next, and making that sale is a lot less costly than acquiring a new customer. There is definitely gold in your transaction data.

But emphasizing marketing to current customers can be difficult if you are addicted to revenue from new customers. One consequence of such an addiction is relative neglect of your existing customers. You may not be jealously guarding and enriching your transaction data. You may not be doing useful customer analytics. You may not even be thanking your customers for their patronage. Ultimately, it is an unsustainable way to build a business.

So how do you kick the habit?

It doesn’t require going to a twelve-step meeting, but neither does it happen overnight. The first step is to establish some benchmarks that accurately describe the state of your business. Next, implement a systematic testing plan. You should be regularly testing your offers, your segmentations, and your targeting.

Third, create custom segmentations of your customer population. Four to seven segments work well, plus you need a marketing plan for each segment. Then build an early warning system to reduce attrition. You want to identify which customers are possible defectors, and which of those are most worth the effort to retain. Finally, design a measurement system that tracks your marketing activities.

When the going gets tough, the tough do customer analytics.

post test power analysis

Mark Klein — Wed, 19 Nov 2008 10:25:11 -0500

As regular readers here know, we’re passionate about testing at Loyalty Builders. It’s the quickest way to improve your marketing campaigns. To spread the testing gospel, we recently posted a free online calculator to tell you how big a sample size is needed for your test, or to compute the power of a test with a given sample size. The calculator is designed for marketers, not statisticians, though several statisticians are using it. To make it more accessible to marketers, we also wrote an extensive tutorial, and that’s where the fun started.

We got to thinking and talking about power calculations made after a test comparing campaigns A and B was completed. [Remember: power is the ability of a test to detect an absolute difference.] Suppose also that the sample response to B is somewhat better than the sample response to A. If the power is high, say the 90% or 95% we regularly urge our clients to use, everyone has confidence in the results of the experiment. Even if the power is merely very good, say 80%, the marketer has learned that his new treatment, used in campaign B, has only a 20% chance that there is actually a real differnece that is not being detected. B is actually that much better than A and there is little value to working to improve B.

But suppose that the calculated power is low, say 40%. Instead of throwing out the experiment as telling us nothing, we realized that there is a 60% chance B could actually yield better results than A. This is a great, counter-intuitive point: if your test concludes no difference, but the power is low, there is actually a fair chance that there is an actual difference that you're missing. You should probably work to improve B or to conduct another, better test.

This conclusion wiped a lot of statistics snobbery off our figurative face. No matter how small your sample size, there may be useful information to be gathered from your test. Sometimes what looks bad is more promising than we thought. Sort of forced us to eat our humble pie [chart].

[My thanks to Dave Meeker and Bill Vorias, two of our math experts who did the work. I was the one who turned his nose up at the low power.]

first time buyers

Mark Klein — Mon, 03 Nov 2008 15:51:25 -0500

Back around 1999, when new products were being introduced as fast as you could put up a website, brand building underwent a quiet revolution.

With markets moving at warp speed there simply wasn’t time to build trust and reputation the traditional way - by advertising. In that super-heated environment the best way to build a brand was to get people to actually try the product. If you could manage to get trial, if the dogs ate the dog food, if they came back to buy again, and if they told their friends, then maybe you had a shot at survival.

The markets haven’t cooled. If anything they’ve heated up. And I still believe that persuading a prospect to try your product or service can create more brand equity faster than all the 30-second Super Bowl spots you can afford. Generating trial is still the best branding strategy I can think of.

The irony is that once a prospect has tried the product – once he’s actually become a new customer – he is virtually ignored because he doesn’t represent much revenue and therefore isn’t part of the elite ‘best customer’ list – even though he represents a terrific brand-building opportunity.

Of course every new customer isn’t a potential ‘best customer’ and finding the good prospects in a new customer list requires analysis. If you’ve been interested in customer analytics you probably know something about RFM (Recency, Frequency, Monetary value) analysis. It’s a helpful way to segment customers, but its not much help analyzing ‘newbies’ because while they’ve bought recently, they don’t have an established frequency of purchase, and the amount they’ve purchased puts them at the bottom of the customer list in terms of revenue.

Fortunately there are some more polished analytics tools (Longbow is one) that go beyond RFM and can help identify new customers who have great promise. All you have to do is be wise enough to recognize their potential and market to them with the care and respect they deserve.

Promising new customers are teetering on the brink of brand loyalty. Your first task is to bring them back for a second purchase. With appealing offers (and a test program to confirm which offers work best) you will build your brand and your revenue at the same time. In effect you’ve created a branding effort that pays its way.

Branding, as practiced by the guru’s, tries to build trust, reputation, and loyalty without any real transactions between sellers and buyers – an ivory tower exercise that burns marketing dollars without assuming any responsibility for revenue. I’ve come to believe it’s a luxury that even rich companies can’t afford these days. Branding by building on a good customer experience is the better way to go.

Retention is the new acquisition

Mark Klein — Sun, 28 Sep 2008 10:04:51 -0400

From here, it looks like the tide is turning.

When you fish the tidal estuaries of our New Hampshire coast, you quickly learn that fish bite when the tide is flowing. Many days I’ve sat in a boat at slack tide, waiting for the current to reverse, the tide to run, and the fish to bite. You watch for those first signs of the water moving, like a leaf or twig and then seaweed beginning to move past your boat. Right now, I am seeing indications that the tide in marketing is changing.

A while ago I wrote in this space (Are your marketing ratios out of whack?) that the marketing spend of most companies was grossly unbalanced, with the bulk of the outlay going to customer acquisition even though most of the revenue is coming from existing customers. However this week at Connections, the user conference of our partner Exact Target, we heard two of the keynote speakers sound a contrary opinion.

Author and blogger Joseph Jaffe spoke the lines in the title of this post, “Retention is the new acquisition.” Then Nigel Travis, CEO of the Papa John’s restaurant chain put it even more bluntly when he said “I don’t need more customers; I need my existing customers to buy more frequently.” We are hearing the same message from our clients—more focus on customer retention and reducing churn.

Of course, as advocates of existing customer marketing, we are encouraged by this trend. We're glad to hear that people are coming around to the place we've been for the last ten years. But we look at retention as more than just stopping customers from defecting, at just holding onto your customers.

Your best new business prospects are often within your customer base – because cross-sell is new business and up-sell is new business. Use mathematical marketing not just to predict which customers are likely to defect, but to learn what to offer them, and to launch up-sell and cross-sell campaigns with double digit response rates. That tide will get stronger and stronger.

Avoiding Mathematical Marketing

Mark Klein — Thu, 11 Sep 2008 10:08:35 -0400

10. Ignorance: They’ve never heard of it.

9. Irrelevance: Most of their customers are one-time buyers, so how could mathematical marketing help?

8. Overconfidence: They think their current marketing program is working fine. They already know who their best customers are. They’re getting 80% of their revenue from 20% of their customers. What could be better?

7. Skepticism: They doubt it works; they believe all customers defect sooner or later.

6. Resistance to change: Regardless of whether it improves the bottom line in the long-run, they are unwilling to change doing business as usual.

5. Inexperience: Mathematical Marketing appears complex, requiring skills and people they don’t have.

4. Cost: It sounds expensive

3. Intimidation: Marketing is an art, and math belongs in the classroom.

2. Myopia: They say it's the customer that drives their business, but it’s really their products.

1. Misdirection: They spend most of their money and time on customer acquisition, which means they believe capturing new customers is more important than selling to current ones.

Loyalty Marketing and the Stockholm syndrome

Mark Klein — Mon, 04 Aug 2008 09:02:27 -0400

Some of my friends have helped me to a better understanding of the difference between loyalty and satisfaction.

They live in a mid-sized city in the mid-West. For a long time their city has been primarily served by Northwest Airlines, so they are frequent fliers on that airline. By many measures they are loyal customers of Northwest.

But when you talk with them you quickly learn that they are more trapped than loyal. They don’t have alternatives, and to them the airline is “Northworst". Loyalty may be high, but satisfaction is not. If there was another carrier, my friends would switch in the blink of an eye.

I’m sure you know instances of customer loyalty being more the lack of alternative vendors than of a favorable opinion. So how do you think a company should operate when many customers who appear to be loyal are really trapped?

It might milk the situation for all it’s worth. This may generate the most near-term revenue, but the backlash could be ferocious when alternative vendors appear, and they will appear. The other extreme is to treat these captive customers as grandly as their frequent patronage warrants. This could lead to a better perception, and may be worth the cost.

But what if there is a third alternative, less costly than the red carpet treatment but yielding the same results? Suppose the Stockholm syndrome applies to these captive customers? (It’s been a long week, and I’m feeling audacious.) The Stockholm syndrome is a psychological response seen in hostages when token acts of kindness by their captors leads to an emotional attachment by the hostage to the hostage taker. Think Patti Hearst. In marketing, the hostage is the trapped customer and the captor is the vendor with extremely large market share.

I think it is entirely possible the syndrome applies to commercial situations. If it does, then token benefits for the captive customers could elicit favorable reactions similar to those from truly loyal customers, for significantly lower cost. Am I being totally cynical, or is this really worth a test?

First of all, a company has to have the corporate will to admit that their loyalty numbers are an illusion. On the positive side, captive customers represent a terrific opportunity. You know who they are, and mathematical marketing gives you a chance to convert them to truly loyal customers at a modest cost. Unfortunately little of what you do will show up in the loyalty metrics until your monopoly is broken, but you can’t sit back and boast about your loyalty numbers until then. You need to understand the dynamics, face them, and build on the advantage you now have. The Stockholm syndrome says your chances are good.

[Thanks to Arthur Einstein for some good ideas on this topic.]

The fallacy behind "break-even point" strategy

Mark Klein — Wed, 23 Jul 2008 14:42:10 -0400

We believe that there is a big fallacy buried in the strategy of companies that use direct mail.
Most direct mailers, especially catalogers, have a ‘house list’ of customers to whom they mail. This list is scored by one methodology or another, for example recency or RFM (Recency, Frequency, and Monetary value). These companies usually talk about “mailing down” the list to some point called the “break-even point”. By this they mean they have ordered their scored list with their best customers at the top and their worst customers at the bottom, and have identified a point on the list below which it is unprofitable to mail. “Unprofitable” typically means that the mailing cost to reach these lower ranking customers exceeds the gross margin dollars that the mailing generates.

Companies who confidently use this approach think they can’t lose money, that it is good business and safe practice to send out more and more catalogs as long as they stay above that sacred break-even point.

Wrong!

Forget for a moment that RFM scoring is notoriously inaccurate in the low and mid-level deciles where the ‘break-even point’ lies. If that’s the method you are married to, rather than more accurate loyalty scoring, then stick with it. To understand the fallacy, it really doesn’t matter what scoring method you use. The bigger problem lies in the customer selection, not in your scoring methodology.

The goal of a campaign is to generate margin dollars—gross profit, not just revenue. The more margin dollars collected, the more successful the campaign. As print mail costs rise, it becomes harder and harder to have a profitable campaign. The obvious alternative is email, but emailing everyone over and over again runs the risk of turning off your customers completely and having them click the unsubscribe link, losing this channel to reach them forever.

Picking who to contact is the key to success.

The break-even strategy is flawed because it doesn’t consider the purchase probabilities for individual customers at the time of the campaign. Marketers that use this strategy calculate aggregate sales, typically by deciles of the scoring metric used, and typically after a campaign is completed. They look at the revenue for each decile and then decide at which decile the gross margin dollars from those aggregated sales drop below the mailing cost. Then they cut off below that decile in the subsequent campaigns.

Instead, customers should be ranked according to their probability of purchasing the products being offered in the next period. Then calculate the predicted revenue from each individual customer to find the set of customers who will generate enough revenue to cover the communications cost. This yields a unique set of customers in the target list. It is the most accurate way to determine when revenue will exceed costs, and the best way to optimize a marketing budget. Building your target list beforehand using purchase probabilities generates more margin dollars than using the break-even point from a previous campaign to determine which customers to touch in a subsequent campaign.

Companies that use the break-even strategy are looking at the problem from the wrong direction. Instead of asking “How deep can I mail and still be profitable?” they should be asking “How profitable can I be with the money I’ve budgeted for the direct mail campaign?”

[My thanks to Bill Vorias of Loyalty Builders for challenging conversations on this topic.]

Left vs. Right: Is Google heading for a fall?

Mark Klein — Mon, 05 May 2008 07:42:17 -0400

Search pages in Google are divided into two parts. On the right are the paid search entries, advertisers who pay Google whenever someone clicks on their ad. On the left is what is commonly called organic search, results turned up by Google’s algorithms for searching the web based on your request. Most people start with organic search, often not going beyond the first five or ten entries. People use the left column results about three times more than the right column ads, according to some estimates. Your mileage may vary.

Strictly speaking, Google makes no money from organic search; those results are merely a platform, an opportunity for Google to post the pay-per-click (PPC) ads in the right column. Almost all of Google’s revenue comes from paid search, the right column ads. It is the primary vehicle for customer acquisition, especially in the SMB space.

However there are signs that the world is changing. More and more companies are realizing that the organic search results in the left column are much more valuable than right column listings. Search Engine Optimization (SEO) is all about setting up your web site to show high in the organic search rankings. It is not easy to do, and a limited number of people know the skills, but more and more companies are trying to learn those skills. As they learn them, they are cutting back their PPC budget and putting the money into SEO. I suspect “SEO for Dummies” will appear soon.

This trend could spell trouble for Google. The company reached its dominant position by solving two problems: 1) how to build a better search engine, and 2) how to monetize that solution. The monetization was in that right column, the PPC ads. Now, as companies switch their emphasis and marketing spend to the left column, Google’s PPC revenue could take a hit. If this hit impacts their R&D spend, some new company could develop a better search engine and reduce their market share, further depressing PPC revenue. There is the possibility of a vicious feedback loop.

There are a few factors working in Google’s favor, however. One is advertiser inertia driven by Google’s enormous market share. A second is the limited number of slots on that precious first page. Even if multitudes of companies figure out how to optimize their sites, there are still only a few that make it onto page one. The rest might have to resort to PPC for visibility.

As for us, Loyalty Builders is shooting for the top of the left column. Look for us there in a few months.

The Devil's Compact in Customer Loyalty Programs

Mark Klein — Sun, 20 Apr 2008 16:08:30 -0400

Sometimes I show some resolve and actually carry out my resolutions. Yesterday was one such day—I cleaned out my wallet, which was too light on money and too thick with cards from various loyalty programs. Pruning those loyalty cards made me look at each program and decide whether participation was worth it, whether there was there a decent quid pro quo for using the program.

The obvious description of customer loyalty programs is that they are a way for a company to thank its regular customers for their continuing patronage with rewards and discounts. Whether it is a coffee card for an extra java after ten purchases or free tickets to Hawaii after a winter of business trip, we all especially enjoy the “free” product or service. It’s nice to be appreciated and thanked.

But there is another side to many loyalty programs, an implicit devil’s bargain: in exchange for the freebies and discounts, we give our contact information (phone, address, email) to our vendor and implicitly give them permission to market to us. Some companies abuse this relationship, bombarding their customers with irrelevant offers and annoying frequency.

But other companies are smarter. Retailers like Vitamin Shoppe, for example, use the programs to track the purchases of customers who would otherwise be anonymous, something ecommerce companies, catalog merchants, and B to B companies take for granted. Done right, we’ve seen retailers make customer associations at point of purchase at rates over 80%. This transaction data is the meat and potatoes of behavioral tracking.

The smart companies make good use of the data they accumulate on the behavior of customers who offer their loyalty card or program number when making a purchase. Companies like Wegman’s Grocers use the techniques of mathematical marketing to predict which customers will buy next, what products they are likely to buy (including those they have not previously purchased), for customer segmentation, and which customers are on the verge of defecting to a competitor. Properly analyzed, a company can learn about more than their customers from this data: there is even valuable product information buried in the transaction data of these customers.

As customers, we tolerate this devil’s compact if our vendor uses what they know about us wisely. If we get relevant offers, if our vendor uses what they learn to talk to us intelligently, they actually build loyalty to their brand. If they don’t play nice, if they abuse the compact, they turn us off and lose our business.

So what kind of loyalty program does your company have? Is it a win-win, of benefit to both the vendor and the customer? Does your loyalty program help you or do you think it is necessary evil that costs you money? Which companies do enough to earn that place in your wallet?

The "Mathematization" of Marketing

Mark Klein — Sun, 10 Feb 2008 11:52:23 -0500

As word spread of my teleconference for Marketing 2.0 (see previous post), an interesting contrarian opinion appeared in my mailbox. Evidently there was something in the meeting notice to cause a reader to suspect that using mathematics to describe behavior could have negative consequences. He wrote “Is it just me, or do others also wonder where this mathematization of social relationships is leading humankind?”

Mathematical marketing

I may be making too much of this comment, but in it I detect an attitude of despair over how the increasing application of math seems to be taking the humanity and flexibility out of human (marketing) relationships. We at Loyalty Builders are definitely not in that camp. We are staunch believers in what some call left brain marketing and others call mathematical marketing. We know that using mathematics can help marketers make a better connection with their customers by, for example, helping them make more relevant offers. Until lately, however, that mathematical approach generally has not been applied with much creativity.

Measuring customer loyalty

Marketers have traditionally tried to understand their customers by studying where they live, their age, their gender, or other demographic factors. But demographics are relatively passive attributes. As you move along the attribute spectrum to ever more active variables, to psychographics and clickstream behavior and ultimately to transaction behavior (what a customer has purchased), mathematics can raise the accuracy of your understanding, the intimacy of your connection to your customer, to ever higher levels. That in turn leads to better customer service, greater efficiencies, and higher profits. Human behavior is not determined by mathematics, but mathematics can be used to measure that behavior.

"We do the math. You do the marketing."

So to answer the question about where this mathematization is leading humankind, I say it is to a place where we can make better use of our capabilities as humans and not to an android society. We recognize and require the creative side of marketing. The right brain is essential for good campaigns, but it will work best when it is operating from a foundation built on solid, left brain analysis. That is the thinking behind our Longbow tag line, “We do the math. You do the marketing.” We practice mathematical marketing to enable your creativity.

Mark Klein is featured speaker at Marketing 2.0 teleconference

Mark Klein — Fri, 08 Feb 2008 16:00:15 -0500

Here's the invitation to the Facebook Marketing 2.0 teleconference where I will be the featured speaker:

You are invited to attend a telephone conference at 1 PM ET on Wednesday February 13 that features Mark Klein talking about behavioral targeting and customer retention. The conference is hosted by Isabel Walcott Hilborn of Marketing 2.0 under the auspices of Facebook. You can see more details, including dial-in information, at http://harvard.facebook.com/event.php?eid=23509056840. If you are not already a Facebook member, you will need to do the simple, free signup.

This event is meant to be conversational, with Isabel asking me questions that she feels are interesting to the participants. I’m sure we’ll be talking about what we’ve learned in the past few years. The formal title is "Customer Conversion and Retention." I'm going to discuss the most common mistakes we see companies making, some surprises we've encountered, and the three biggest things we've learned. I hope you will join us.

Behavioral Targeting vs. Demographics

Mark Klein — Tue, 22 Jan 2008 08:12:49 -0500

In the current post on Metrics Insider, Josh Chasin argues that behavioral targeting is not the be-all and end-all, that demographics is alive and well. He calls demographics “the targeting concept that wouldn’t die.” He says further that “more than half of U.S. ad dollars are placed against traditional demographic targets.”

If behavioral targeting marketers, the ones who focus on clicks, are missing opportunities by ignoring demographics, think how much more they are missing by ignoring the true behavioral indicator, purchase behavior. Imagine a scale of customer behavior that runs from active on one end to passive on the other. At the active end is a customer’s purchase behavior. At the passive end are their demographics. In between are their clicks from browsing activities. The clicks are active, but not as committed as a purchase. Remember the old story about bacon and eggs: the chicken is involved but the pig is committed.

We know from several years of analytics experience that the closer you get to the active end of the scale, the more accurate are the predictions. The clicks and demographics are valuable, useful pieces, but past purchase behavior is still the best foundation for offers to existing customers.

Old School vs. New School Testing

Mark Klein — Thu, 10 Jan 2008 08:15:21 -0500

David Baker has a good column in Email Insider this week, What Really Makes An Email Program Click. In it, he writes "Remember one thing: You test to prove or disprove your hypothesis, not just for the exercise."

To me, that sounds like old school testing, where one ‘thing’, one hypothesis, is tested. Formally, that is called split-run or A/B testing. There is only one variable, and two variations of it. It works, but it is time consuming and (if there is more than one hypothesis) needs lots of test subjects and money. Consequently many companies just don’t do it, relying instead on their intuition.

Today however you can test many variables simultaneously, and even uncover the interactions between the variables. The two main types of marketing tests are testing offers and testing segments. Examples of offer tests are HTML vs. text, blue background vs. red, $10 off on $100 purchase vs. $20 off on $200, etc. Examples of segment tests are recent purchasers, potential defectors, some geography, various loyalty deciles, etc. The math is more complicated (it’s called Factorial Design), but there are vendors to do it for you, and the results are outstanding.

So yes, absolutely test. An email test is generally over within two days, and it is one of the two best ways to improve your marketing campaigns. But test everything, not just ‘your hypothesis.’

Can one metric do the job for ranking customers?

Mark Klein — Sat, 05 Jan 2008 08:28:13 -0500

I am admittedly biased. When asked what I think is the best metric for ranking customers, I invariably recommend our proprietary Loyalty Score for a two reasons. First, it is highly predictive, corresponding to the probability that the scored customer will make a purchase in the next period. Second, it’s tested and proven through use by many clients over the past few years. We have a lot of confidence in it. Besides, I enjoy tweaking people who think the only good measure is revenue.

Lately however you might hear a different answer from me. It’s not that Loyalty Score is less useful or has deteriorated. The opposite is true: our algorithms are better and Loyalty Score is more powerful than ever. Rather, it is my perspective, my point of view, that is changing. Now that the new tools are in place, I find myself advocating a different strategy to our clients. Instead of merely ranking the customers on a linear scale, I recommend exploring the whole customer set to find customer segments with varying needs, and then marketing to them with appropriate messages. Identifying those segments usually requires looking at more than one metric, so I’m reluctant to offer up a single answer to the ‘what’s the best?’ question.

For example, if you are planning a win-back campaign, you’ll want to rank the customer both on Recency and Revenue and Loyalty Score. You need Recency to find the customers that haven’t bought in a while; Revenue to find the customers who are worthwhile, and Loyalty Score to figure out which ones you have the best chance of recapturing.

Another example is finding customers who are good candidates for cross-sell. Here I would rank customers based on their category score (how many different types of products have they purchased) and the change in their Loyalty Score. I’d pick candidates whose Loyalty Score is changing in a positive direction (they are good customers), but whose category score is low (they are buying deeply but not broadly).

Other possible metrics are Risk Score (likelihood of defection) and an RFM score. What is your favorite? Will one metric do the job? Should I go back to my simple-minded point of view?

Have you had a customer conversation lately?

Arthur Einstein — Thu, 29 Nov 2007 09:45:15 -0500

I like the elegant language we use talking about customer loyalty -‘predictive analytics’ and ‘behavioral targeting’ come immediately to mind. They make loyalty work sound almost like an academic discipline. Good for the ego.

On the other hand, the 50-cent words tend mask the beauty of what customer loyalty applications really accomplish - which is to is let a business person have a privileged and private ‘conversation’ with each and every customer – no matter how many customers are on the books.

You need to be a good listener to learn anything from the conversation. And the process involves some sophisticated math, but the computer does that. However getting to know your customers and their habits definitely isn’t brain surgery.

Listening starts with learning from your accounting databases, looking at orders, invoices, shipping information and other ordinary transaction data. This puts the conversation on a purely factual basis - because the information on an invoice is not an opinion. It’s what’s actually happened.

In a lot of organizations these facts are still touched only buy billing clerks. But upload the facts to a customer loyalty application, and that application can reveal a lot more useful information about what customers have been up to lately. All you have to do is listen.

Customers’ data will tell you what they’ve been buying, and re-ordering. Buying patterns emerge. Over time it’s easy to see if the patterns are changing and to spot buying trends. A good listener will be able to forecast customers’ future purchases with accuracy – which can help plan inventory needs.

The conversation can predict when customers should buy next, and what they’ll buy. It can identify the customers who should have bought but didn’t and what they’re not buying that others like them are. The customers will also tell you loud and clear when they’re unhappy enough that they just might leave and take their business elsewhere. Since you’re talking with every customer, about every item they’ve purchased, this kind of conversation produces more dependable answers than a few face-to-face interviews.

Of course, listening isn’t worth much unless you do something about what you hear – but we don’t call that analytics. We call that marketing.