Borrowed Brilliance: Three Industries Destination Marketers Should Be Studying Right Now

Earlier this year, I argued that destination marketers need to stop looking exclusively at other destination marketers for inspiration in my first Borrowed Brilliance blog. Not because the work happening across our industry isn’t good, but because when everyone is attending the same conferences, seeing the same case studies, and benchmarking the same campaigns, we shouldn’t be surprised when destinations all start launching similar ideas.

If we want different results, we need different inputs.

That idea became the foundation for my U.S. Travel Educational Seminar for Tourism Organizations (ESTO) presentation this year, where I shared three examples of industries solving many of the same marketing challenges DMOs face every day: different industries and audiences, but many of the same customer behaviors and marketing challenges.

Here are three lessons from other industries that can help destination marketers approach familiar marketing challenges with a fresh perspective.

Real Estate: Market Long Before Someone Is Ready to Choose

Vacation booking and business event decisions aren’t made overnight. Neither are home purchases.

Take Teravalis, a 37,000-acre master-planned community outside Phoenix that had to sell a vision before it had much to show. Buyers weren’t purchasing an existing neighborhood; they were buying into a vision of what the community would become. That meant marketing couldn’t focus on immediate conversions. Much like destination marketing, it had to build awareness, pique curiosity, and create an emotional connection to an experience people couldn’t have yet. They understood that the buyer cycle is long and messy, and the marketing had to speak to them at every stage.

So we designed the marketing approach around the entire buyer journey. Storytelling introduced the long-term vision, and earned media and PR supported awareness hurdles. Paid media stayed in front of potential buyers as they researched, with retargeting keeping the community top of mind during the long research and consideration stages. We segmented audiences based on behavior in their CRM, and the messaging and creative shifted for every stage of the buyer journey. Success wasn’t measured only by homes sold; it was measured by engagement signals. We measured repeat visits, content engagement, growing remarketing audiences, and better-informed prospects entering the sales process. Marketing success looked for signals that people were moving closer to a decision, not just the end result. 

When it came time to promote the Grand Opening, the campaign wasn’t starting from scratch. Months of storytelling, awareness-building, and audience nurturing had already introduced people to the vision and created anticipation for what was coming. The Grand Opening campaign generated 110 million impressions across Google, Meta, programmatic, and out-of-home advertising with over 80,000 clicks, 782 qualified leads, and more than 2,000 event attendees.

But the real story isn’t the campaign metrics. It’s what made those results possible. Those leads weren’t generated by a single campaign but the culmination of months spent building awareness, sparking curiosity, and creating demand long before people were ready to buy. By the time the Grand Opening arrived, there was enough momentum to generate nearly a quarter of all qualified leads for the second half of the year during the launch period alone.

That’s what happens when marketing influences the entire customer journey instead of focusing only on the final conversion. 

The lesson for destination marketers is simple: Most travelers don’t wake up one morning and book a trip to your destination. 

They compare different places. They ask friends. They search for things to do. They build itineraries. They shop for flights, hotels, and activities. They come back to your website multiple times before making a decision.

Yet many tourism marketers still focus on the booking moment, especially when budgets get tight, or stakeholder pressure focuses just on the final metrics. 

The takeaway for destination marketers is to look beyond the booking moment. Don’t let your marketing strategies and media plans be overly influenced by reports that focus only on the final conversion. That’s what most analytics platforms are designed to measure, but it’s only one part of the story. The traveler decision-making process isn’t a straight line. People discover, compare, validate, revisit, and eventually decide.

Those interactions may not be conversions, but they’re often what make conversions possible.

Healthcare: Expertise Doesn’t Matter If People Can’t Find It

If there’s one area where we want to trust the experts, it’s healthcare. And successful healthcare organizations have become incredibly good at one thing: answering questions.

Patients rarely begin their search by looking for a provider’s name. They’re searching for answers to very specific problems: “Adult ADHD testing near me,” “Why does my shoulder hurt?” or “Do I need physical therapy after knee surgery?” The organizations that show up aren’t always the ones with the best doctors. They’re the ones that have packaged their expertise around the questions people are already asking.

One healthcare organization we partnered with had exceptional specialists and expanding services, but much of that expertise wasn’t appearing when prospective patients searched online. Their website described their services well enough, but it wasn’t organized around how patients actually searched for care. It also wasn’t helping filter out patients who weren’t a good fit, creating unnecessary friction for both prospective patients and providers. Our strategy was simple: make their expertise discoverable while helping the right patients find the right care.

Instead of simply producing more content, we restructured the entire content strategy around real search behavior. We expanded high-intent service and condition pages around the questions people needed answered, optimized provider bios to be more discoverable, and created educational content that connected expert knowledge with real search intent. Every element of the strategy, from site architecture and internal linking to localized SEO and generative engine optimization (GEO) efforts, worked together to help the organization show up wherever patients were searching, whether that was Google, AI search, or other answer engines.

The impact went far beyond increased website traffic. Organic sessions grew more than 200% year over year, organic form submissions increased nearly eightfold, and organic search became one of the organization’s primary patient acquisition channels, accounting for 42% of bookings.

The lesson for destination marketers is that your destination already has valuable expertise. Your locals know the hidden gems. Your sales team knows the questions meeting planners ask before submitting an RFP. Your visitor services staff answers the same traveler questions every day. The challenge isn’t creating expertise. It’s making that expertise discoverable.

The easier your expertise is to find, the more likely your destination is to become part of the consideration set before a competitor does.

Retail: Stop Waiting Until the Campaign Ends to Learn

Retail marketers have something I wish more destination marketers embraced. They’re less concerned with defending a campaign and more interested in improving performance. Rather than relying on a single creative concept or a fixed set of assets, they treat every campaign as an opportunity to test, learn, optimize, and scale what works.

One retailer we partnered with, Blue Waters, had built a loyal community and strong engagement on social media, but they needed more than views and likes. They needed sales. As TikTok’s algorithm evolved and audience behavior shifted, the challenge wasn’t creating more content. It was turning social engagement into a scalable ecommerce channel.

Instead of treating TikTok like another social platform, we approached it like a storefront. We launched TikTok Shop campaigns and tested dozens of short-form creative concepts, starting with single-product videos to quickly identify what resonated. Real-time performance data informed every decision, from creative and messaging to media allocation. As winning content emerged, we scaled it, retargeted shoppers, and supported live shopping events that kept customers moving from discovery to purchase.

The results demonstrated the value of that approach. TikTok Shop revenue increased 125% month over month, purchase rates climbed 87%, cost per purchase dropped 72%, and overall revenue grew 35% in just two quarters. More importantly, TikTok evolved from an awareness channel into one of the brand’s highest-performing sales channels.

The takeaway for destination marketers is that optimization shouldn’t wait until the campaign is over. Many DMOs may consider optimizing media throughout a campaign, but they’re often working from a single creative concept or one set of campaign assets. Retail marketers think differently. They build testing into the creative strategy itself, continually evaluating different messages, visuals, formats, and audiences to see what resonates.

That means refreshing creative regularly, promoting the organic content that’s already proving it works, shifting media dollars toward what’s performing, and scaling winners across channels. Instead of asking, “How is this campaign performing?” they’re asking, “What can we learn today that will make it perform even better tomorrow?”

That level of continuous optimization takes time and expertise, which is why having a performance-focused agency partner can make such a difference. The right partner is continuously testing creative, identifying new opportunities, reallocating budget, and refining campaigns in real time. That not only saves your team valuable time, it also creates a more efficient media strategy, improves return on ad spend (ROAS), and frees your staff to focus on the bigger-picture initiatives that move your destination forward.

Looking Outside the Tourism Bubble

These examples aren’t really about real estate, healthcare, or retail. They’re about marketing.

Real estate reminds us that great marketing doesn’t begin at the moment of conversion. It builds awareness, sparks desire, and nurtures confidence throughout a long and often messy decision-making process.

Healthcare demonstrates that expertise only creates value if people can find it. The best content is built around how consumers search and structured to be discovered across traditional search, answer engines, and AI.

Retail proves that testing beats assumptions every time. Success comes from continuously optimizing not only your media, but also your messaging, creative, formats, and audiences while campaigns are still live.

The industries may be different, but human behavior isn’t. That’s why Borrowed Brilliance isn’t about copying someone else’s campaign. It’s about recognizing great marketing wherever it happens and asking one simple question: How would this work for my destination?

The best ideas in destination marketing probably won’t come from the next DMO case study. They’ll come from looking beyond our industry, borrowing the best ideas, and adapting them to the unique challenges we solve every day.

Where to Find Marketing Inspiration

The Borrowed Brilliance concept doesn’t stop with this blog or my ESTO presentation. Some of the best ideas in destination marketing are hiding in industries most of us aren’t paying attention to. Here are four simple ways to bring more outside perspective into your work.

Curate your feed.

Fill your LinkedIn, Instagram, and newsletter subscriptions with marketers from outside the destination industry. Follow retail, ecommerce, healthcare, real estate, higher education, hospitality, and tech. The goal isn’t to copy what they’re doing. It’s to understand how they’re solving marketing challenges that look surprisingly similar to yours.

Use AI as your research assistant.

Instead of asking AI for tourism marketing ideas, ask what tourism can learn from other industries. Try prompts like, “What can destination marketers learn from retail?” or “How do universities market during long decision-making journeys?” AI is a powerful tool for uncovering patterns, strategies, and examples you might never have found on your own.

Follow marketing events.

You don’t have to attend Cannes, Adobe Summit, HubSpot INBOUND, Google Marketing Live, or Shoptalk to benefit from them. Most conferences publish keynote recordings, session recaps, and trend reports after the event. Spend an hour reviewing the biggest takeaways and ask yourself, “How could this apply to my destination?”

Choose partners who bring outside perspective.

The same principle applies to the partners you work with. Some of the strongest agency relationships combine deep destination marketing expertise with experience across other industries, bringing fresh ideas and proven strategies that might otherwise never make their way into tourism. The best partners don’t just understand destinations. They bring outside perspective that helps destinations stay ahead.

The best destination marketers go beyond studying destination marketing. They study great marketing wherever they find it. And sometimes, the idea that transforms your next campaign comes from an industry that has nothing to do with tourism. 

Follow along for my upcoming Borrowed Brilliance blogs, or catch insights on LinkedIn and emails from Go Fish Tourism + Business Events.

That uncertainty is reasonable. Ad costs on the platforms most eCommerce brands rely on have risen steadily, and setting budgets based on last year’s benchmarks are going to pay for that lag. New channels emerge regularly, each promising the kind of early-mover returns that compress quickly once the auction fills up. What worked eighteen months ago may be quietly underperforming today, and the reporting that should surface that often doesn’t.

This piece lays out a practical framework for building and maintaining an eCommerce paid media budget grounded in where channel economics actually stand right now.

Key Takeaways

How Much Should You Actually Be Spending?

Before touching channel percentages, the total budget needs to be grounded in something more defensible than last year’s number plus inflation.

Most profitable eCommerce brands invest 10 to 20% of gross revenue in marketing. Early-stage brands in the $500K to $2M revenue range often need to lean higher, reaching toward 18 to 25% to build acquisition momentum. Mature brands at $10M and above can typically pull back to 8 to 12% as existing customers and organic channels carry more weight. Trying to grow aggressively on 5% rarely works, but spending 30% without the economics to support it can burn cash faster than it builds customers.

For context: Gartner’s 2026 CMO Spend Survey puts average marketing budgets at 7.8% of company revenue across all industries. That figure reflects large enterprises across many sectors, many of them in efficiency mode. For eCommerce brands actively growing, it’s a floor, not a target.

The right total budget is anchored to three numbers: the cost to acquire a new customer, the value of that customer over time, and how long the business can sustain the gap between those two figures before it closes. If acquisition costs are rising faster than customer value, more spend won’t solve the problem. If the math is strong and returns are stable, the constraint on growth is usually underinvestment rather than poor channel selection.

The 70/20/10 Framework: What It Is and Where It Came From

The 70/20/10 allocation framework has been a staple of budget planning across industries for decades. At Google, Eric Schmidt and Jonathan Rosenberg formalized the ratio in their book How Google Works: 70% of resources on the core business, 20% on emerging products built on that foundation, and 10% on completely new ideas. Gmail and Google News both emerged from that 10%.

Applied to paid media, the logic translates directly: protect what’s working, invest in what’s promising, and reserve a small pool for what might matter next. The framework endures because it addresses two ways budget allocation tends to go wrong. Teams that put everything into proven channels feel safe in the short run but stagnate because channels saturate, costs rise, and a brand that never tests anything new is always in reaction mode. Teams that spread resources across too many unproven channels never invest enough in any one of them to learn what works. The 70/20/10 split enforces discipline in both directions.

For eCommerce specifically, the three buckets work as follows:

One important caveat: the 70/20/10 split describes how to weight your budget, not a prescription for which specific channels belong where. A channel in the 20% bucket for one brand may be a 70% channel for another, depending on category, audience, and what the data shows.

How Allocation Changes by Revenue Stage

The 70/20/10 framework is a useful structure, but what belongs in each bucket shifts meaningfully as a brand grows. A single allocation applied across every stage produces the wrong answer at most of them.

Early stage ($0 to $2M revenue). Concentrate your budget. Early-stage brands should put the majority of paid spend on one prospecting channel. The goal at this stage is finding buyers, testing creative, and establishing whether costs can reach profitability. Dividing budget across multiple channels at this stage slows learning without meaningfully reducing risk. Meta broad targeting provides the fastest feedback loop for most product categories. TikTok works better for trend-driven products with audiences skewing younger. Pick the channel that maps to your category and customer, and don’t expand until you have at least 60 days of consistent performance at scale.

Scaling stage ($2M to $10M revenue). Diversify deliberately. A practical starting split for most scaling brands: 40% toward Google Shopping, 25% toward Meta, 15% toward Microsoft, 10% toward retargeting, 10% toward testing. The logic here is that Meta or TikTok builds demand, and Google Shopping captures the purchase intent that demand-creation spending produces. The channels reinforce each other when funded in the right ratio. At this stage, email and SMS should also be earning a meaningful budget line because email consistently delivers among the highest returns of any marketing channel and functions as a natural counterweight to rising acquisition costs on paid platforms.

Mature stage ($10M+ revenue). Protect the core and shift more toward retention. Mature brands typically move toward 40 to 50% paid acquisition, 25 to 30% retention investment, 15 to 20% creative production, and 5 to 10% experimentation. At this stage, the economics of keeping an existing customer rarely look worse than acquiring a new one. The paid media budget increasingly serves to protect market position and expand reach deliberately, rather than to build the customer base from scratch.

The Channel Landscape Right Now

Allocation frameworks only work when the underlying channel economics are current. A few things have shifted enough recently to affect what the buckets should include.

According to Triple Whale’s analysis of nearly 35,000 eCommerce brands, Meta CPM rose 20% year over year in 2025, reaching a median of $14.19  and every industry in the dataset saw costs increase. Google Ads costs are moving in the same direction. Triple Whale found that Google Ads median CPA rose 12% year over year in 2025, with CPM up 10% to $12.79.

Costs are rising on the platforms where most eCommerce brands concentrate spend. That doesn’t mean those platforms stop working, but it does mean the 70% bucket needs to be earning its keep more rigorously than it did two years ago.

Go Fish Digital’s own data across 300+ clients reinforces the channel shift already underway. According to our eCommerce media trend analysis, spending has increased across all major platforms over the past four years: Amazon up an average of 36%, Google up 30%, YouTube up 32%, and Meta up 13%. The Amazon growth is notable because as more consumers start product searches directly on the platform, retail media has moved from a line item brands experimented with to one that matters for anyone selling physical goods.

Amazon’s 2025 performance data tells a useful story, too. CPA fell, ROAS improved, and conversion rate strengthened, but amidst those positive signals, CPM surged 47% year over year, the sharpest increase of any metric in Triple Whale’s Amazon benchmarks. More brands competing for the same high-intent audience is pushing costs up even on a platform where efficiency has historically been strong. Understanding where costs are rising fastest, and where returns still justify the price, is the input the allocation framework needs to work properly.

For a broader view of where enterprise budgets are moving across search, retail media, social, and AI discovery, see our analysis: Where Enterprise Retail CMOs Should Move Budget.

The Budget Line Most Brands Get Wrong: Creative

There’s no universally agreed-upon benchmark for how much of a paid media budget should go toward creative production. The figures that circulate (10% is the target, most brands are spending 3%) are practitioner rules of thumb more than documented standards. What the data does support clearly is the underlying logic behind them. Motion’s 2026 eCommerce benchmarks found that creative accounts for 70% of campaign performance variance on Meta—more than audience targeting, bidding, or budget allocation combined. If creative is doing that much of the work, funding it like a line item rather than a production cost is the more defensible approach.

The production gap reinforces the point. The median eCommerce brand produces 8 new ad creatives per week. Top-quartile performers produce 22 or more, using batch production, modular creative systems, and AI-assisted variation to keep volume high without proportionally increasing cost. That difference in output is partly a creative strategy question, but it’s also a budget one, and brands that treat creative production as an afterthought to the media plan tend to find out the hard way when fatigue sets in and performance drops without an obvious explanation.

This matters because creative is the primary lever in paid social. Think of it this way: the media budget rents attention, but the creative is what does something with it. On platforms like Meta and TikTok, the algorithm routes spend toward ads that earn engagement and drive action. Better creative is a cost efficiency improvement because stronger ads generate more return per dollar of media spend. Underfunding creative means renting attention you’re not making the most of.

Meta now favors accounts running 15 or more active creatives for optimal ad delivery, a threshold that requires a production system, not occasional campaign assets. An eCommerce brand running the right channel allocation with underfunded creative is leaving performance on the table at every stage of the purchase process.

At Go Fish, we treat performance creative as a core component of paid media strategy. We have a production system built to generate, test, and iterate on creative at the volume and pace the platforms require. See more of how that works in practice: TikTok Ads Strategy for Performance Marketing: A Creative-First Playbook and How to Test and Scale Paid Social Ad Creative.

When to Reallocate and How Often

The most common budget failure in eCommerce paid media is failing to adjust when channel economics change. The right reallocation cadence is quarterly review. Daily reactive shifts erode the data needed to understand what’s working. Annual reviews are too infrequent as channel costs move month to month. A useful rule of thumb: if any single channel’s cost to acquire a customer moves 25% or more in either direction over a 30-day window, that’s worth investigating as a potential structural shift rather than a blip, and may warrant an off-cycle reallocation decision.

Seasonality is the other reallocation driver that gets underweighted. Ecommerce brands with significant fourth-quarter concentration need a budget calendar that front-loads creative testing earlier in the year so campaigns entering the holiday period are already optimized rather than still learning. Q4 typically represents 42% of annual eCommerce revenue and CPMs during that period run 35 to 45% higher than annual baseline. Waiting until October to ramp holiday spend means paying peak prices for data you could have collected at lower cost in July.

The 70/20/10 framework was not designed to remain static all year. Today’s 20% growth bet may earn a spot in the 70% core. Last quarter’s core channel may be showing signs of diminishing return and need a reduction. No allocation framework should hold rigid across a full year of shifting channel economics.

A Starting Framework by Budget Size

For brands formalizing paid media allocation for the first time, a simplified starting point by monthly budget:

Under $10K/month: Concentrate 100% in one or two channels, typically Google Shopping and one paid social platform, until you have 60 or more days of consistent performance data. Don’t divide budget before you have enough in any single channel to act on what you’re learning.

$10K to $30K/month: Apply a rough 70/20/10. Seventy percent to your proven acquisition channel or channels, 20% to a secondary channel showing early results, 10% to creative testing or a new platform. Reserve at least 10 to 15% of total budget for creative production separate from media spend.

$30K to $100K/month: Full 70/20/10 across a multi-channel mix: Google Shopping, Meta, email and SMS retention, one or two growth channels (TikTok, retail media, YouTube). Creative budget should scale with media spend. Quarterly reallocation reviews become essential at this level, and a measurement infrastructure including attribution modeling, channel-level reporting  is worth investing in.

$100K+/month: The framework holds but the analysis becomes more sophisticated. Understanding what’s actually driving performance versus what’s capturing credit requires more rigorous measurement: incrementality testing and media mix modeling both become more relevant at this level. This is where guessing is most expensive and data quality matters most.

Not sure whether your current paid media allocation matches your stage and goals? Go Fish Digital’s paid media team audits eCommerce accounts across search, social, and social commerce and builds the budget framework from there. Request a paid media audit.

Frequently Asked Questions

What percentage of revenue should an eCommerce brand spend on paid media? Most eCommerce brands in active growth invest 10 to 20% of gross revenue on total marketing, with paid media representing the largest portion of that. Early-stage brands often need to be toward the higher end of that range to build momentum. Mature brands with strong organic and retention channels can operate toward the lower end. The right number depends on your margins, your cost to acquire customers, and what that customer is worth over time — not on what the industry average says.

What is the 70/20/10 rule in paid media? It’s a budget allocation framework that divides spend into three buckets: 70% toward channels with a documented track record of positive returns, 20% toward channels showing early promise that haven’t yet been fully optimized, and 10% toward experiments with new platforms or formats. The framework was popularized by Google and has been widely adopted in marketing because it balances protecting what works with creating space to find what might work next. The percentages are a starting point, not a fixed rule. What belongs in each bucket should reflect actual performance data.

How often should I review my paid media budget allocation? Quarterly. Annual reviews are too slow given how quickly channel costs move. Daily changes erode the data needed to understand performance. Reviewing every quarter gives enough time for changes to register in the data while still allowing meaningful course corrections across the year.

What channels should be in the 70% “proven” bucket for eCommerce brands? For most eCommerce brands at scale, the 70% bucket typically includes Google Shopping or Performance Max (for capturing existing purchase intent), Meta (for prospecting and retargeting), and email and SMS (for keeping existing customers engaged). The exact mix depends on your category, audience, and what your own data shows is working. A channel belongs in the 70% bucket when it has delivered consistent positive returns for at least two consecutive quarters with enough volume to draw reliable conclusions.

How much should I budget for creative production? At minimum, 10% of total paid media spend. Most brands allocate far less than that and pay for it in weaker ad performance. Creative production should be treated as a media efficiency investment: better assets drive better returns from the same media spend, because platforms like Meta and TikTok distribute ads based partly on how well they perform, which means creative quality directly affects cost.

When should I start testing new channels? Once your primary channel has delivered consistent, profitable returns for at least 60 days and your total monthly budget exceeds roughly $10,000. Below that threshold, dividing budget across multiple channels typically produces insufficient data in any of them to make informed decisions. When you do expand, fund the new channel meaningfully because a token allocation generates token data.

Is it worth advertising on TikTok Shop? For the right product and category, yes. TikTok’s median CPM in 2025 was $13.26, which is still below Meta’s $14.19, and the platform’s in-feed shopping format compresses the path from discovery to purchase in ways traditional eCommerce cannot match. Products that demonstrate their value visually, solve a specific problem, and land in the $20 to $60 price range tend to perform best. See how Go Fish clients have approached this: How Brands Are Driving Revenue with TikTok Shop.

How do I know when my channel allocation is working? The right metrics depend on your stage, but the most reliable indicator is whether the cost to acquire a new customer is stable or improving relative to what that customer is worth over time. Platform-level metrics like return on ad spend and click-through rate are useful for diagnosing specific campaigns, but they don’t tell the full business story. If acquisition costs are rising and customer value isn’t keeping pace, the allocation may be working at the campaign level and failing at the business level.

Are you putting your paid media budget in the right places?

We can help you evaluate your current channel mix, identify where spend may be underperforming, and determine where there’s room to reallocate or scale.

Someone might search Google for “best running shoes for flat feet.” In ChatGPT, they might explain that they’re training for their first half marathon, run mostly on pavement, want to spend less than $150, and found their last pair too heavy.

That conversation gives ChatGPT more context about what the person is actually trying to solve.

There’s a lot more information in that conversation for an advertising platform to work with.

OpenAI began testing ads in ChatGPT in the U.S. in February 2026. Since then, the company has continued developing ChatGPT Ads, including tools for advertisers to create campaigns and measure performance.

ChatGPT Ads put paid media into a place where people are already researching products, asking follow-up questions, comparing options, and working through decisions. Marketers now have to figure out where that fits within the customer journey and whether the additional conversational context can make advertising more relevant.

Here’s what we know so far.

How Do ChatGPT Ads Work?

ChatGPT Ads are sponsored placements that can appear below ChatGPT responses for eligible users. Ads are selected based on factors including conversational context, advertiser inputs, creative, and landing-page relevance.

According to OpenAI’s documentation, advertising operates separately from ChatGPT’s answers. Advertisers can’t pay to influence which brands or products ChatGPT recommends organically.

So it’s possible for a brand to show up in a ChatGPT response without advertising. It’s also possible to advertise alongside a conversation without being mentioned in the answer itself.

OpenAI says its ad system can consider several signals when deciding which eligible ad to show:

Advertisers can also provide context hints describing conversations, topics, or keywords where their product or service may be useful. OpenAI notes that these aren’t exact-match keywords and don’t guarantee placement within a specific conversation.

For teams used to organizing paid search campaigns around keywords, this requires some adjustment.

How Are ChatGPT Ads Different From Search Ads?

The biggest difference is the amount of conversational context available around a user’s intent. Paid search often begins with an individual query, while ChatGPT can draw on the details someone shares throughout a conversation.

Think about someone shopping for a new mattress. Their Google history over a few days might look something like this:

A ChatGPT user could provide most of that information in one conversation. They might explain how they sleep, what’s wrong with their current mattress, how much they want to spend, what their partner needs, and which brands they’ve already ruled out.

The intent isn’t necessarily stronger. There’s simply more context available around it.

OpenAI has described this as an opportunity to reach people while they explore, compare, and decide. For advertisers, the interesting question is how useful that additional context becomes for targeting and performance.

ChatGPT Ads vs. Search Ads vs. Organic AI Visibility

The comparison is useful because ChatGPT now creates room for paid and organic visibility within the same customer experience.

Context Hints Add Another Layer to Targeting

One of the more interesting features in ChatGPT Ads is context hints.

Context hints allow advertisers to describe the conversations, topics, or keywords where their offering may be relevant. They aren’t exact-match keywords and don’t guarantee an ad will appear in a particular conversation.

This is where customer research becomes particularly useful. Search data can tell you what people type into Google. Sales calls, customer interviews, reviews, support tickets, community discussions, and on-site behavior can help fill in the circumstances around those searches. These sources are useful for understanding customer language and decision-making, regardless of whether an AI platform currently cites them frequently.

Ad Creative Has to Make Sense in the Conversation

OpenAI’s creative guidance for ChatGPT Ads recommends clear, specific messaging that explains what a product offers, who it’s for, and when it might be useful.

That guidance fits the environment.

If someone has spent several minutes explaining exactly what they need, an ad that says “Discover the Difference” isn’t adding much.

An ad that clearly explains why a particular product or service fits the situation has a better chance of being useful.

OpenAI also recommends creating distinct ad variations rather than making minor changes to the same message. Different creative can address different needs, use cases, or customer situations.

The destination matters too. Sending every ChatGPT Ads click to a homepage throws away much of the relevance that got the click in the first place. Product pages, category pages, comparison content, FAQs, buying guides, service pages, and other specific resources may make better landing pages depending on the conversation.

How Are ChatGPT Ads Bought and Measured?

ChatGPT Ads currently support CPC and CPM buying models, with campaigns optimized around either clicks or reach. Advertisers can measure impressions, clicks, spend, CTR, average CPC, average CPM, and conversions.

OpenAI says its system uses a relevance-weighted, second-price auction when selecting among eligible advertisements.

Advertisers can also add UTM parameters to destination URLs and track ChatGPT Ads traffic in their existing analytics platforms.

Most of this will look familiar to performance marketers. What we don’t have yet are mature benchmarks for how ChatGPT traffic behaves across industries and use cases.

A $5 click from ChatGPT and a $5 click from Google aren’t automatically equivalent. The user may be at a different point in the buying process, and the behavior after the click may look different as a result.

For early campaigns, we’d pay attention to conversion rate, lead or revenue quality, landing-page engagement, assisted conversions, and incremental reach alongside the standard platform metrics.

Paid and Organic AI Visibility Need to Be Measured Separately

Paid AI visibility is exposure purchased through advertising, while organic AI visibility comes from a brand being mentioned, cited, surfaced, or recommended within an AI-generated response.

OpenAI says advertising doesn’t influence ChatGPT’s answers. Paying for ChatGPT Ads won’t make a brand more likely to be recommended in an organic response.

A company might be mentioned or cited in the answer because ChatGPT considers it relevant. It might appear as a sponsored placement. In some conversations, it could do both.

For brands already investing in generative engine optimization (GEO), this distinction matters for measurement.

If branded visibility increases after launching ChatGPT Ads, teams need to know how much came from paid placements and how much came from organic mentions, citations, or recommendations.

Establishing an organic AI visibility baseline before testing ads can make that much easier to understand.

ChatGPT Doesn’t Fit Neatly Into One Marketing Team

There’s another practical issue companies will have to work through: ownership.

SEO teams typically manage organic search. Paid media teams manage search advertising. Content teams build educational and decision-stage resources. Brand teams manage positioning. Analytics teams measure what happens across all of them.

A ChatGPT conversation can touch several of those functions at once.

Someone can ask a question, receive information pulled from third-party sources, encounter brands within the answer, compare products, and see an ad without ever leaving the conversation.

Internally, those activities may sit with several different teams. The customer sees one experience.

Companies testing AI advertising will probably need more coordination between paid media, SEO, content, analytics, and brand than they’re used to with a traditional paid search campaign.

How We’d Approach an Early ChatGPT Ads Test

The performance data available today is still early and isn’t category-specific enough to make broad budget decisions. Running a controlled test against your own benchmarks will tell you more about how ChatGPT Ads perform for your business

The first round of testing should help marketers understand where ChatGPT shows up in their customers’ buying process. Testing is also more accessible now that OpenAI has reduced the minimum budget requirement to $0, removing the need for a large upfront investment. That gives advertisers more room to start small, compare performance against existing channels, and learn where ChatGPT fits in their media mix.

ChatGPT Ads FAQs

Can advertisers pay to appear in ChatGPT answers?

No. OpenAI says advertisements operate separately from ChatGPT’s answers. Paying for ChatGPT Ads doesn’t allow an advertiser to influence which brands, products, or services ChatGPT mentions or recommends organically.

Do ChatGPT Ads use keywords?

Advertisers can provide topics, keywords, and context hints to help OpenAI understand where an ad may be relevant. Context hints aren’t exact-match keywords and don’t guarantee placement within a specific conversation.

How much do ChatGPT Ads cost?

ChatGPT Ads support CPC and CPM buying models. Actual costs depend on the advertising auction and campaign. Advertisers should establish their own performance benchmarks as more campaign data becomes available.

Should brands use ChatGPT Ads or invest in GEO?

ChatGPT Ads and GEO serve different purposes. Ads provide paid exposure within ChatGPT, while GEO focuses on improving a brand’s organic visibility across AI-generated answers, citations, mentions, and recommendations. For many brands, the more useful approach will be measuring and testing both rather than treating them as interchangeable.

Where ChatGPT Ads Fit From Here

ChatGPT advertising is still early. Capabilities will change, more performance data will become available, and marketers will get a better sense of which products and customer journeys are a good fit.

There’s no reason to overhaul a paid media strategy because ChatGPT Ads exist. There is a good reason to start learning how customers use conversational AI when they’re researching your category.

That includes looking at the questions they ask, the comparisons they make, the sources AI platforms rely on, where your brand already appears organically, and which conversations have enough commercial intent to warrant a paid test.

OpenAI maintains a collection of ChatGPT Ads resources for advertisers as the platform develops.

For marketers testing the channel, the useful benchmark won’t be whether ChatGPT Ads can replace Google Ads. It’ll be whether they reach customers in relevant moments that your existing media mix isn’t reaching as effectively.

Considering a ChatGPT Ads test?
Start with the right use case, benchmarks, and measurement plan. Talk to our team about where ChatGPT Ads could fit into your media mix.

To illustrate the point, we ran a simple test. We asked Google’s AI Mode, “What snack keeps your milk from getting lonely?” No brand name, no category hint; we just typed the question the way a person would naturally think it. It came back with a generic answer about cookies, then went on to extol the virtues of chocolate chip cookies. No OREO in sight, just a lonely glass of milk.

Then we tried the same test with another iconic brand association: “What candy melts in your mouth, not in your hands?” ChatGPT immediately connected the phrase to M&M’s.

Same experiment. One hit, one miss.

That inconsistency is the story and the challenge brands face in a digital age. AI brand visibility isn’t something you either have or don’t. AI visibility exists on a spectrum, and most brands have no idea where they fall.

AI Brand Visibility Is Not the Same as Brand Equity

Brands like OREO Cookie have spent decades building powerful brand associations through advertising, packaging, and pop culture. “Milk’s favorite cookie” is a prime example. Consumers readily connect the phrase with OREO Cookies, making it a strong example of traditional brand equity.

Despite that strong brand equity, Google’s AI Mode didn’t make the same connection.

AI doesn’t experience your brand the way people do. It reads signals.If your strongest brand associations live in TV spots and consumer memory rather than in clean, consistent, crawlable text across the web, you may have a visibility gap you don’t know about.

Great marketing is written for its audience. It is emotional, memorable, and built around a campaign moment. The OREO Cookies example illustrates this well. “Milk’s favorite cookie” is a brilliant line. It earns attention, drives recall, and holds up across decades.

The problem is that AI doesn’t experience campaigns. It reads online text.

The same tagline that wins awards for a human audience may leave almost no structured signal for AI to connect back to a brand name. That isn’t a failure of the marketing team. It’s a gap in how the brand’s digital presence has been built around the marketing.

Closing that gap means adding a parallel layer: plain, consistent, crawlable language that states exactly what your brand does, who you serve, and why it matters. No metaphor required.

Why AI Gets Your Brand Wrong

AI tools don’t look up your brand in a master database. They assemble a picture from whatever content is most consistent and credible across the online sources they can access, whether it’s your website, directory listings, review pages like Yelp, press coverage, forums like Reddit, Wikipedia, and more.

When those sources agree, AI answers confidently. When they conflict, grow stale, or simply don’t clearly reinforce your core positioning, AI fills in the blanks on its own. It may describe what you do in outdated terms, confuse you with a competitor, omit your strongest differentiators, or identify the category but not the brand.

This isn’t a glitch. AI can only work with the brand information it finds across the sources it uses.

Common reasons brands end up misrepresented:

How To Audit Your Brand in AI Search

One way to find out where you stand is to do a manual brand audit. It doesn’t take very long, and the results can be quite surprising.

Open Google AI Mode, ChatGPT, Gemini, and Perplexity separately. Use the same prompts on each platform and keep the wording consistent. Try questions like:

Document what each platform says. Note where they agree, where they differ, and where they’re wrong. That gap between what AI says and what you know to be true is your baseline. It tells you what you’re working with before you start fixing anything.

One important note on methodology: ask the way a human would, not the way a traditional search query would. Don’t guide the AI toward the right answer by giving it category hints. If it needs that level of prompting to get your brand right, that’s a gap worth knowing about.

How to Fix AI Brand Visibility Gaps

The good news is that most of these brand association gaps are fixable. The bad news is that this isn’t a one-time project; it’s an ongoing maintenance effort, closer to technical SEO than a campaign.

One thing worth understanding before you start: AI doesn’t take your word for it alone. What you say about your brand on your own site is only one input. Third-party sources can carry more weight because they’re independent. Reviews, press mentions, directories, and forum discussions aren’t just good for SEO. They’re the sources AI trusts most when forming its picture of who you are. That’s why so many of the fixes below focus on what others say about you, not just what you say about yourself.

Here’s the playbook:

A word about Reddit 

Reddit deserves a special call-out here. It’s one of the most heavily cited sources in AI-generated answers because AI treats it as unfiltered, unsponsored consumer opinion.

As a brand looking for AI recognition, the instinct may be to jump in and start posting, but that approach tends to backfire. Reddit communities are skeptical of brand presence, and anything that reads as promotional can do more harm than good. 

The real value is in listening and observing. Search your brand name and your category. 

That intelligence tells you exactly where your off-site narrative has gaps, and you address those gaps through better content and earned mentions elsewhere, not by showing up in the thread yourself.

How To Know If Your AI Visibility Strategy Is Working

Re-run your original prompt set monthly. Keep a simple log of what each platform says and track whether responses shift over time. You don’t need expensive software to start.

For traffic signals, set up a GA4 segment for AI-referred visits (referral sources from ChatGPT, Perplexity, and similar platforms are increasingly trackable). Watch whether that traffic grows as you improve your off-site presence and content.

If you reach the point where you want to track this at scale, across many prompts, competitors, and platforms, tools like Semrush and seoClarity are building out AI brand monitoring features worth evaluating. But the manual version works well enough to establish a baseline and build your instincts before committing to a platform.

The Bigger Picture

If OREO Cookies, one of the most culturally embedded brands on the planet, can be a ghost in AI search, no brand is too well-known to have this problem.

The brands that audit and fix this now are building an advantage that compounds. AI is increasingly the first stop in the buyer journey, not a secondary reference. What AI says about you, or doesn’t say, shapes perception before a prospect ever lands on your site.

The brand equity you’ve built over years can be quietly undermined by what AI can’t find, can’t verify, or can’t clearly attribute to you. The fix starts with knowing where you stand.

Go Fish Digital helps brands build visibility across traditional and AI search. If you want to know what AI is saying about your brand, you’ll want to request a free AI Audit. 

This article turns a core slice of our on‑demand workshop, The GEO Gap: How to Diagnose What’s Hurting Your AI Search Visibility, into a practical self-serve process. You’ll learn how to test whether AI actually recommends your brand, score the answers with a simple framework, find out why competitors get cited instead, and measure your AI Share of Voice so you can track progress over time.

Key Takeaways

Part 1: Workshop — See How AI Treats Your Brand

Step 1: Start with a clean conversation

Open ChatGPT, Gemini, Claude, Perplexity, or another AI assistant.

Use temporary chat, private mode, or a new account session where possible. This limits the influence of your previous conversations, brand history, and personalization.

You want the model’s view of the market, not its view of you.

Step 2: Ask the broad category question

Use this prompt and record whether your brand appears:

“What are the best [products or services in your category]?”

Examples:

Step 3: Add the real customer problem

Now ask a follow-up using the same category, but add the type of detail a real buyer would include:

“We are/I am a [type of customer] dealing with [specific problem or constraint]. What [product or service] should we/I look at?”

Example:

“We are a club soccer team working with a limited budget. Which companies should we consider for printing our uniforms?”

The list may change once the assistant has to solve a specific problem. That change is the point of the exercise.

Step 4: Score the answer

Do not stop at whether your brand appeared. Review the answer using these five checks:

1. Do you appear?  Is your brand included in the answer at all?

2. Are you picked? Is your brand actually recommended, or is it only one name in a longer list?

3. Do you get the link? Does the mention include a clickable link or citation to your website?

4. Do you win the head-to-head? Choose one competitor from the answer and ask:

“Compare [your brand] and [competitor]. Which one would you recommend for [specific customer need]?”

5. Are you described correctly? Check whether the answer accurately describes your products, services, audience, pricing, features, and positioning. Flag anything outdated, incomplete, or incorrect.

Step 5: Repeat the test

One answer is not enough to establish a pattern. AI responses can change between platforms and between runs.

Repeat the test:

Track which brands appear, which ones get recommended, whether they receive a link, and how they are described.

You are not trying to find a fixed AI ranking. You are looking for patterns in where your brand appears, where it disappears, and where competitors are consistently winning.

Part 2: Compare — See Why the Cited Source Won

If Part 1 showed that a competitor was recommended instead of you, received the link, or appeared when your brand did not, the next step is to understand why.

This diagnostic assumes the baseline AI-readiness work, including site structure, technical setup, and digital PR, is already done. If it isn’t, start with our AI Performance Review Checklist first.

Before you dive in, confirm your page is actually eligible to be cited: make sure it loads normally, isn’t hidden behind a login, and has a real webpage version, not just a PDF. If you’re not sure how to check this, ask your web team for help.

Then identify the page AI cited instead of yours. Open it side by side with your own, and work through the checks below.

Freshness matters too: 

Do This at Scale

Go Fish’s AI-powered marketing intelligence platform, Barracuda, runs this type of diagnostic at scale. Barracuda’s Content Similarity module breaks a page into sections, scores how well each one matches what someone is actually asking, and flags precisely where a “ranking” page is falling short, no manual checklist required.

Ready to go beyond the individual tests in this article? Our free GEO Scorecard provides a deeper look at how your brand performs across AI search and helps surface the visibility gaps that could be holding back mentions, citations, and recommendations.

A potential customer can ask ChatGPT which providers to consider or see Google explain a topic in an AI Overview. Those answers may use a brand’s own pages, but they can also draw from news coverage, trade publications, reviews, forums, and other third-party sources.[2][5][9]

This is where Digital PR and generative engine optimization, or GEO, connect. Digital PR creates data-driven stories and earns coverage beyond a brand’s website. GEO focuses on how a brand appears in AI-generated answers. That coverage gives AI systems more independent context about what the brand does and what it is known for.[6][7][8]

One placement will not guarantee an AI mention or citation. But research from Muck Rack and BuzzStream shows that earned media is a meaningful part of the source mix, especially when people use AI to compare options. [2][5]

The question I hear most often is not, ‘Does Digital PR still work?’ It is, ‘Does it still matter if search is changing this fast?’ My answer is yes – because the need for third-party credibility has not gone away.

 Key Findings 

What Is Digital PR? 

Digital PR is an earned media and link-building strategy built around stories publishers want to cover. Campaigns often use surveys, public datasets, or a brand’s own data to uncover a useful trend, ranking, or comparison.[10] 

A Digital PR service handles the research, analysis, content, and media outreach. Coverage can earn backlinks, referral traffic, and brand mentions. Those outcomes support SEO and give GEO more credible context beyond the brand’s own site.[6][8][9]

Why Digital PR Matters More in AI Search 

AI answers draw from more than a brand’s website 

AI tools can use information from news sites, trade publications, review pages, forums, and other sources across the web. That means a brand’s visibility can depend partly on what credible third parties say, not just what it publishes itself.[2][3][5]

The exact share varies by study and prompt type. Muck Rack and BuzzStream used different methods, but both found that earned sources make up a substantial part of AI citations.[2][5]

The source mix also changes by AI platform and prompt. BuzzStream found that most citations in its 595-prompt study were unique to a single platform. This is why Digital PR can help brands earn coverage across a wider mix of relevant sites instead of relying on one publication to reach every AI platform.[4]

We have seen that variability in our own AI visibility research, too. In a July 2026 study of Go Fish Digital across ChatGPT, Google AI Overviews, and Perplexity, we analyzed 1,554 responses across 73 prompt wordings and found that results shifted substantially depending on both the platform and how the question was framed. We also saw that AI systems can rewrite, shorten, or expand a user’s original question before retrieving information, which means the sources surfaced for one prompt may look very different from those surfaced for another.

People use AI to compare brands 

AI is not only used for quick facts. People can ask which brand fits a need, how two companies differ, or which product is a better fit. Google describes AI Mode as useful for complex comparisons.[9]

For questions closer to a buying decision, AI tools often cite third-party content. In BuzzStream’s study, earned content made up nearly all citations for evaluative prompts. A brand’s website can explain what it does. Independent coverage gives AI tools another source to evaluate those claims.[5]

AI tools lean on journalism for industry trends 

AI tools use different sources depending on the question. Muck Rack found that industry-trend questions cited journalism at more than twice the rate of how-to questions. [2]

That gives Digital PR a natural role in AI search. Timely research, expert commentary, and data-driven campaigns can earn coverage as a story develops, giving AI tools more credible reporting to draw from when explaining what is changing in a market. [2]

That is why I do not think the AI shift changes the case for Digital PR. It changes the way we explain the value, but the underlying job is still the same: earn outside proof that makes a brand easier to trust.

How a Digital PR Service Supports AI Search Visibility 

Digital PR supports the part of GEO that happens beyond a brand’s own website. It creates useful information, earns independent coverage, and helps connect the brand with the topics it wants to be known for.[6][7][10] 

 1. Digital PR creates original research worth covering 

Strong campaigns give journalists something new to work with, such as a survey result, stat, ranking, or trend. The Digital PR team handles the methodology, analysis, writing, visuals, and outreach needed to turn that research into a credible story that earns coverage.[7][10]

Voctos identifies original data as a strong Digital PR asset for GEO. In a BuzzStream interview, Kevin Indig said proprietary research can win citations, but stressed that distribution is critical. The data needs a clear story and a plan to reach the right publishers.[7][10]

The campaigns that tend to perform best are rarely the ones that stay closest to the brand. More often, it is the tangential ideas, the ones just outside the core product or service, that create the strongest coverage opportunities. They feel timely, they have a better shot at fitting the news cycle, and they do not come across like an ad for the company. 

At the end of the day, earned media only works when we stop asking, ‘Do we like this idea?’ and start asking, ‘Would a journalist actually care about this?

 2. Digital PR earns independent coverage around the right topics 

A brand can explain its expertise on its own website. Digital PR adds outside context by earning coverage from publishers that readers and AI systems may already use.[2][3][7]

A strong campaign relates to the brand’s industry without reading like an ad. When relevant publications cover the research, the brand appears alongside the topic it wants to be known for. That can give AI systems more third-party context about the brand.[6][7][8]

 3. Digital PR builds a broader, more relevant source mix 

Not every AI platform cites the same pages, and the biggest outlet is not always the most useful one. Muck Rack found journalism citations across more than 20,000 outlets and notes that niche trade publications can be especially useful for B2B questions.[2][3]

A Digital PR service builds the media list around the audience, topic, and campaign. The goal is relevant coverage across a mix of trusted sources, not a single popular news site.[3][4]

 4. Digital PR connects PR, SEO, and GEO 

One piece of coverage can earn a backlink, mention the brand in a relevant context, and later appear as an AI source. [6][8]

Google says the same SEO fundamentals still apply to AI Overviews and AI Mode. BuzzStream’s topical authority interview also emphasizes the need for content, internal links, backlinks, and brand mentions to reinforce the same topics. That connection also shows up in the execution of Digital PR. In our own pitch testing, we have seen how choices like anchor text can influence journalist engagement while also shaping SEO and how a story may be interpreted in AI summaries. Digital PR is most useful when it reinforces what a brand is already building through SEO and owned content. Digital PR is most useful when it reinforces what a brand is already building through SEO and owned content.[9][11]

Digital PR cannot guarantee inclusion in an AI answer. It can build a stronger third-party footprint that gives a brand more chances to be discovered, mentioned, or cited.[2][4][5]

 FAQs About Digital PR Services and AI Search 

What does a Digital PR service do? 

A Digital PR service develops campaign ideas, creates the research and content behind them, pitches relevant journalists, and reports on the coverage the work earns, including backlinks and brand mentions.

Can Digital PR help a brand appear in AI search? 

Yes. It can support AI visibility by earning credible coverage, links, and mentions across sources AI tools may use. Results vary by platform, question, industry, timing, and access, so no service can guarantee an AI citation. 

Do backlinks still matter in an AI search world? 

Yes. Backlinks still support SEO and can send referral traffic. For GEO, they should be considered alongside brand mentions and AI citations.

Is Digital PR the same as GEO? 

No. GEO is the broader work of improving how a brand appears in AI-generated answers. Digital PR supports that work by earning third-party coverage, links, and mentions outside the brand’s website. Owned content, technical SEO, and site accessibility still matter.

Do I still need Digital PR for AI search visibility?

Yes. AI search has changed where people get their answers, but it has not changed the value of credible third-party coverage. Digital PR helps brands earn that coverage through original research, timely stories, and relevant publications. That gives people and AI tools more independent proof to work with.

Do I need both GEO and Digital PR? 

Usually, yes. GEO helps a brand show up well in AI-generated answers, while Digital PR helps create the outside proof that supports that visibility. GEO is the broader strategy. Digital PR is one of the strongest ways to support it because it earns mentions, links, and coverage from other sites.

Build the Third-Party Proof That Supports AI Search Visibility 

AI search has not replaced the need for good website content or strong SEO. It has added another layer: what the wider web says about your brand.

Digital PR helps build that layer through original research and earned coverage. The result is not a guaranteed AI citation. It is a stronger, more credible presence across the sources people and AI systems use to understand a category. Learn how Go Fish Digital’s Digital PR services can help build that earned media footprint.

 Sources 

[1] Atomic AGI. “Digital PR for AI Search: The Complete Strategy Guide for 2026” April 2, 2026.

[2] Muck Rack. “Earned Media Still Drives 84% of AI Citations. Here’s What That Means for PR.” May 7, 2026.

[3] Muck Rack. “What Are AI Citations? How PR Teams Can Track and Earn Them” July 21, 2026.

[4] BuzzStream. “50 AI Statistics Digital PR and SEO Should Know (2026)” June 30, 2026.

[5] BuzzStream. “What Kind of Content Does AI Cite (Based on Prompt Type)?” April 9, 2026.

[6] BuzzStream. “How AI Mentions and Cites Your Brand (New Study)” July 14, 2026.

[7] Voctos. “Digital PR for GEO: How Brand Mentions Earn AI Citations” July 21, 2026.

[8] Ahrefs. “Top Brand Visibility Factors in ChatGPT, AI Mode, and AI Overviews (75k Brands Studied)” December 12, 2025.

[9] Google Search Central. “AI Features and Your Website” updated December 10, 2025.

[10] BuzzStream. “Kevin Indig on Citations vs Mentions and Proprietary Research” July 22, 2026.

[11] BuzzStream. “Building Your Website’s Topical Authority with Kevin Indig” October 27, 2025.

Behind each of those outcomes is the same pattern: TikTok’s For You Page (FYP) puts a creator demo, before-and-after, or haul in front of shoppers at the exact moment they are ready to buy. Meta, Pinterest, and YouTube Shorts are building similar feed-first discovery systems, so what works on TikTok can also inform the rest of your social commerce mix.

Here’s how to win the algorithm, make your products findable, and turn feed attention into revenue. Go TikTok-first, then extend each play to Meta, Pinterest, and YouTube.

Key Takeaways for Social Commerce Discovery

How TikTok’s Algorithm Drives Product Discovery

TikTok’s FYP ranks proof that a video is worth someone’s time: watch time, completion rate, replays, shares, and saves are the strongest signals, whether the interest is a passing glance or an active recommendation.

TikTok parses audio, on-screen text, and captions to understand a video’s subject and match it to the right viewers. Say the product type and benefit out loud and show it on screen. That naming convention helps the algorithm index new or niche products that don’t yet have an audience.

TikTok Shop deepens the loop. When a shoppable video drives detailed views, add-to-carts, or purchases, that commerce performance becomes another signal the content is useful, and high-performing product videos can expand reach for both the SKU and the creative.

Creative testing is non-negotiable. Test hooks (the first two to three seconds), lengths (quick cuts under 15 seconds versus longer 30-45-second explainers), formats (before/after, tutorials, GRWM, hauls), and talent (creator versus brand voice). Once you know what wins on TikTok, check the same pattern against Instagram Reels and YouTube Shorts to see how much carries over. TikTok’s result still decides what ships, since its feed gives you the clearest signal fastest.

Make Your Products Findable With Social SEO

Users type natural phrases, product types, problems, and benefits, into each platform’s search bar, and the algorithms also index captions everywhere, alt text on Instagram and Pinterest, and audio and on-screen text specifically on TikTok and YouTube.

Write captions and on-screen text that sound human but still carry search context: product type, core benefit, and use case. Say and show the product name, like “vitamin C brightening serum for dull skin,” instead of a vague line like “you need this.”

Native shop and catalog data have to work as hard as your captions do. Titles and descriptions should be concise, keyword-rich, and free of fluff, with variants, materials, sizes, and care details that people actually search. This helps on-platform search and improves catalog ad matching.

Build the phrase strategy on TikTok first, mixing high-intent phrases (“how to,” “best,” “for [problem]”) with conversational terms. Then adapt: align keywords to titles and descriptions on Instagram and YouTube, write save-friendly, benefit-led descriptions on Pinterest, and add alt text everywhere for accessibility and visual search.

Turn Creators and UGC Into Your Discovery Engine

Social shoppers look for real people who resemble them to validate fit, quality, and results, and a steady drumbeat of UGC (unboxings, reviews, routines, before/afters) multiplies the discovery doors your products can walk through. People trust people before they trust product pages.

Build a TikTok-first creator and affiliate program that rewards performance. Blend fixed fees for production with tiered bonuses tied to discovery KPIs (saves, shares), product detail views, and TikTok Shop conversions. Give creators clear briefs on product naming, benefits, and hooks, while preserving their style.

Incentivize UGC with seeded product, surprise-and-delight kits, and social-only bundles. Provide shot lists built around quick transformations, side-by-side comparisons, and “I wish I knew this” tips, in formats that work in the feed: GRWMs, hauls, duets and stitches, challenges, and Q&As.

Once an asset wins on TikTok, syndicate it to Instagram, Pinterest, and YouTube Shorts. Tag products natively wherever possible and retarget viewers with catalog ads, so the engagement a creator generates on one platform raises the ceiling on organic distribution everywhere else.

Tighten the Funnel With Native Shops and Seamless Checkout

Every extra click after discovery risks losing the sale. Forcing users to leave TikTok or Instagram to complete checkout introduces load time, login hurdles, and context-switching that erode intent, especially on mobile.

Enable native shops where your audience is most active: TikTok Shop first, then Instagram Checkout. Keep catalogs clean, prices accurate, and inventory synced, so natively tagged products shorten the path from feed to order and let the algorithm learn from commerce outcomes.

Invest in catalog quality: strong thumb-stopping imagery and short video, clear titles, variant details, and benefit-forward descriptions. Pair with dynamic catalog ads to resurface items viewers engaged with but didn’t buy, and set retargeting windows to match your product’s consideration cycle.

Test offers inside native shops: limited drops, bundles that solve a complete use case, timed promos, and creator-exclusive codes. Measure conversion rate, but also watch how each offer moves upstream discovery signals like saves and shares.

Running TikTok Shop campaigns that aren’t converting to sales? Go Fish Digital’s social commerce services are built for exactly this, from feed strategy to native checkout.

Design Shoppable Video and Live Experiences That Spark Discovery

On TikTok, native-feeling, story-driven videos outperform polished ads for discovery. Keep the arc tight: problem, product in action, payoff. The brand stays present but secondary to the utility or entertainment. Authenticity and pacing beat polish.

Make shopping effortless with product tags, stickers, and overlays so viewers can tap directly from content to details. Place tags near the reveal, and restate the product name and benefit both verbally and on screen.

Use live shopping to collapse education and purchase into one session. Mix formats: drops and limited releases, deep-dive demos, expert Q&A, before/after showcases, and flash bundles, with a moderator handling chat, pinning products, and surfacing FAQs in real time.

Clip the best live moments into short videos for TikTok, then reformat creator segments for Instagram Reels, YouTube Shorts, and Pinterest so new viewers can discover them outside the original stream.

Measure, Test, and Keep Your Discovery Strategy Compounding

Anchor dashboards to early- and mid-funnel indicators that predict sales.

Instrument attribution for native checkout. TikTok Shop and Instagram Checkout don’t map one to one with traditional analytics, so rely on platform reporting, unique product IDs, and creator or promo codes that trace back to specific videos, creators, and offers.

You keep earning TikTok’s algorithm, one video at a time.

Frequently Asked Questions About Social Commerce Discovery

What are the key social commerce product discovery trends for 2026?

Zero-search feeds on TikTok, with Meta, Pinterest, and YouTube following the same pattern, shoppable short-form video and live shopping, and native in-app checkout are defining discovery in 2026. Brands that publish feed-native content, keep catalogs clean, and enable native checkout capture more first-touch moments.

How do I optimize product discovery on TikTok using its algorithm signals? 

Prioritize watch time, completion, replays, shares, saves, and shoppable actions. Name products and benefits clearly in audio, on-screen text, and captions. Enable TikTok Shop so strong commerce performance amplifies reach, and keep testing hooks, lengths, formats, and creator versus brand voice.

How do I make my products findable with social SEO across platforms? 

Start with natural, descriptive captions and on-screen text on TikTok that name the product type, use case, and benefit. Keep catalog titles and descriptions concise and keyword-rich, then adapt the same approach for Instagram and YouTube titles, Pinterest descriptions, and alt text everywhere.

Do I need a different strategy for Meta, Pinterest, and YouTube, or can I reuse TikTok content? 

Prove the format on TikTok first, since its feed rewards fast iteration. Then reformat winning assets: trim for Reels and Shorts, adapt captions to each platform’s search behavior, and write save-friendly descriptions for Pinterest.

What is social commerce product discovery?

Social commerce product discovery is the process of finding products directly through social platforms rather than starting with a traditional search engine or retailer. On TikTok, Instagram, Pinterest, and YouTube, discovery can happen through recommendation feeds, creator content, social search, shoppable video, and live shopping. 

More on Social Commerce From Go Fish Digital

That is because there is no single best LLM for every SEO use case. ChatGPT, Claude, Gemini, and Perplexity can all support SEO strategy, content creation, keyword research, technical SEO analysis, and competitor research, but their strengths vary by task. In many cases, the quality of the data, prompt, and verification process matters more than the model itself.

This guide compares the leading LLMs across five common SEO applications and explains where model choice makes a meaningful difference. It also covers the larger factors that affect output quality, including access to real SEO data, human review, and whether you are optimizing for traditional search or AI-generated results.

One clarification before we start. “Which LLM is best for SEO” is really two questions. The first is which assistant your team should use to do the work. The second is which AI surfaces your clients need to appear in. Those have different answers, and mixing them up is why so much AI SEO advice feels muddled. This guide handles both, but separately.

The Best LLM for Each SEO Task

Why Popularity Does Not Determine the Best LLM for SEO

Popularity is a poor way to choose an LLM for SEO because usage data measures consumer adoption, not performance across SEO tasks.

By web-visit share, Momentic’s July 2026 analysis of Similarweb data put ChatGPT at 53.9% worldwide in May 2026, Gemini at 27.9%, and Claude at 9.2%, with a tighter spread in the US. By monthly active users, Sensor Tower’s State of AI Report via TechCrunch put ChatGPT above 1.1 billion, Gemini at 662 million, and Claude at 245 million.

Same market, two metrics, meaningfully different pictures. Neither measures fitness for SEO work, which is consumer behavior data being asked to answer a professional tooling question. That is also why Perplexity is in this guide despite being the smallest of the four by traffic: it earns its slot on one capability, live retrieval with visible citations, not on scale.

What Matters More Than Choosing an AI Model for SEO

1. Whether the LLM Has Access to Real SEO Data

Access to reliable data matters more than model choice for most SEO tasks.

Every model will invent search volumes, difficulty scores, and traffic estimates with complete confidence. Strategy decks are where that does the most damage, because a fabricated figure looks exactly like a real one sitting inside an otherwise sound recommendation. In Keyword.com’s State of AI and Automation in SEO 2026, 70% of respondents named poor output quality, hallucinations, or quality control time as their biggest limitation.

Worth noting on that survey: the sample is a rank-tracking platform’s own audience, so it skews technical. Read the direction as reliable and the precise figures as indicative.

The fix is a workflow rule. Feed verified exports or connect live data through an API, then force traceability in the prompt itself.

2. Whether Someone Verifies the AI Output

Only 1% of respondents in the same survey described their SEO work as fully automated. Practitioners were least willing to automate tasks where errors carry the highest cost, including content writing at 58%, link building at 51%, and technical SEO at 40%.

That caution is justified. A weak AI-generated draft creates editing work. A faulty redirect rule, canonical recommendation, or technical SEO fix can affect live revenue pages. LLMs are most useful when they accelerate analysis and execution without replacing expert review. 

3. Whether You Are Optimizing for the Right AI Search Surface

This one is easy to get wrong because the chat products are the visible part of the market.

At Google I/O 2026, Sundar Pichai put AI Overviews at more than 2.5 billion monthly active users and AI Mode at over 1 billion, roughly a year after launch.

Now compare the chat products. Conductor’s 2026 AEO/GEO Benchmarks Report, built on 13,770 enterprise domains and 3.3 billion sessions, found AI referral traffic accounts for 1.08% of all website traffic, growing about 1% month over month, with 87.4% of it coming from ChatGPT. Search Engine Land has the industry breakdown. Note that Conductor’s dataset is enterprise-skewed and single-vendor, so treat the direction as firm and individual numbers as directional.

Google’s AI surfaces are where visibility is won. The chat products are mostly where your team works, with ChatGPT as the one that also sends meaningful traffic.

Get those three right and model choice becomes a preference. Get them wrong and no model saves you.

Limitations of Using LLMs for SEO

Before we get started, there are a few factors all teams need to take into consideration:

Best LLMs for SEO by Task

The prompts below are starting patterns from working practice, not tested recipes. Adapt them to your stack and check the output before you trust it. And remember, the same prompt can yield different results based on your history. 

1. Best LLM for SEO Strategy Development

Recommendation: whichever frontier model you already pay for.

Strategy work means holding large, messy inputs at once: a keyword export, a crawl, a year of Search Console data, and a competitor set. All four handle this class of task, and the differences between them are small relative to the difference between grounded and ungrounded input.

Force traceability rather than shopping for a smarter model:

You have been given a keyword export as a CSV.

Rules:

Now cluster these keywords by search intent and flag the five clusters

with the best ratio of combined volume to average difficulty.

The third rule is the one most teams skip, and it is the one that surfaces what your data is missing.

2. Best LLM for Keyword Research

Recommendation: Perplexity or Gemini to diagnose. None of them to measure.

Here the model does matter, for one reason. Diagnosis means working out why a competitor gets cited when you do not, and that needs live retrieval with visible sources. ChatGPT and Claude will answer from training data unless you force a search, which is how teams end up presenting stale competitor sets as current.

Measurement is different. Asking a model whether it recommends a brand samples one response from a probabilistic system, on one platform, at one moment. Purpose-built visibility trackers exist because that sampling problem is real.

Search the live web. For the query “[client query]”, identify:

Do not summarize the topic. I only want the citation set and why.

Two findings should shape your GEO recommendations more than most advice does.

Earned media dominates. Muck Rack’s May 2026 “What Is AI Reading?” study, covering more than 25 million links cited by ChatGPT, Claude, and Gemini across 17 industries, found earned media accounts for 84% of all AI citations. Paid and advertorial content accounts for 0.3%. Journalism alone contributes 27%. That figure has held between 82% and 89% across three editions since July 2025, which matters more than the exact number. Digital PR is GEO work now.

Schema is oversold as a citation lever. Ahrefs tracked 1,885 pages that added JSON-LD between August 2025 and March 2026, matched them against 4,000 control pages, and ran a difference-in-differences analysis. AI Mode moved +2.4% and ChatGPT +2.2%, both indistinguishable from noise. AI Overviews moved -4.6%, small but statistically significant. Search Engine Journal covers the methodology.

Read that narrowly, because the industry is already overstating it. Every page in the study was already heavily cited before schema was added. It shows that adding schema to pages AI already cites does not get them cited more. It does not test whether schema helps a page break in for the first time, and Ahrefs says so. Keep schema for rich results and machine parsing. Stop selling it as a citation strategy.

3. Best LLM for Technical SEO Analysis

Recommendation: ChatGPT or Claude. They trade the top spot, and the gap between them is small. 

These two are the only real candidates here. Gemini and Perplexity are strong elsewhere in this guide, and on private-codebase coding evaluations Gemini sits well behind both, for reasons covered below.

Three things make the choice between ChatGPT and Claude less consequential than it looks.

First, the benchmark everyone quotes is compromised. SWE-bench Verified remains the most-cited coding benchmark, but OpenAI withdrew from it in February 2026 over contamination concerns.

Second, the scores that circulate are not comparable to each other. Vendor-published figures are usually produced on the vendor’s own agent scaffolding, while independent evaluators like Vals AI run their own harness. Change the harness and the ordering at the top changes with it. Any guide that hands you a confident ranking is reading vendor numbers as though they were neutral ones.

Third, and most relevant to SEO work: performance on public code does not predict performance on code the model has never seen.

Scale AI runs SWE-bench Pro as two leaderboards through identical scaffolding. The public set is 731 tasks from open-source repositories. The private set is 276 tasks drawn from 18 proprietary startup codebases the models have never encountered.

The ranking flips between them. On the public set, GPT-5.4 leads the three major assistants at 59.1%, ahead of Claude Opus 4.6 at 51.9% and Gemini 3.1 Pro at 46.1%. On the private set, Claude Opus 4.6 leads at 47.1%, GPT-5.4 falls to 43.4%, and Gemini 3.1 Pro drops to 32.2%.

Read that carefully. Claude gave up 4.8 points moving to unseen code. Gemini gave up 13.9. GPT-5.4 gave up 15.7, the steepest fall of the three, despite topping the public leaderboard.

Two caveats keep this honest. Confidence intervals on the private set run to roughly 6 points, so treat Claude and ChatGPT as tied there rather than ranked. And these are specific tested versions, so read the pattern rather than the leaderboard positions.

Why it matters here: your scripting work runs against client repositories, crawl exports, and log files that no model has trained on. The private set is the closer analogue to the job, and it is the number almost nobody quotes.

Both points lead the same direction: the models are close, and the benchmarks do not measure this work anyway. SWE-bench tests resolving GitHub issues in Python repositories. No public benchmark tests whether a model can write a redirect map that will not break your canonical structure.

So spend the effort on verification. Two rules catch more errors than model selection ever will.

Make it explain before it writes. Errors surface in prose that hide in regex.

I need a redirect rule. Before writing any regex:

Wait for my confirmation, then write the rule.

Test against a real sample every time. Run generated logic against a crawl export before it touches a server. Treat model output as a first draft from a capable engineer who has never seen the site.

4. Best LLM for SEO Content Creation and Optimization

Recommendation: any of the four. The differentiator is not prose quality.

Drafts do not fail because a model writes badly. They fail because it holds your house rules for three sections, drifts back to generic AI cadence, and the editor catches it on the fourth pass instead of the first.

That is a systems problem. Encode the rules explicitly, load them where they persist across every draft, and check output mechanically rather than by feel:

Style rules, applied to every draft without exception:

After writing, output a compliance table: each rule, pass or fail,

and quote any line that violates one.

The last instruction is what makes it stick. A model that has to audit itself catches most of its own drift.

5. Best LLM for SEO Research and Competitor Analysis

Recommendation: Perplexity, once your data handling is sorted.

The Tow Center for Digital Journalism at Columbia ran 1,600 queries across eight AI search tools, feeding each a verbatim excerpt from a news article and asking it to name the headline, publisher, date, and URL. The excerpts were chosen so that pasting them into Google returned the original in the top three results.

The tools answered incorrectly more than 60% of the time. Perplexity had the lowest error rate at 37%. Grok 3 was worst at 94%. Gemini produced a completely correct response on one occasion.

Read that properly, because it is not a product endorsement. Perplexity won, and Perplexity was still wrong on more than a third of queries where the answer was trivially findable in Google. The case for it is not reliability. It is that Perplexity shows its sources inline, so checking its work costs seconds instead of a separate search. For competitor scans, prospect research, and verifying a claim before it reaches a client deck, that verification speed is the actual product.

Two findings from that study deserve their own line.

Paying more made things worse. Perplexity Pro and Grok 3 each answered more prompts correctly than their free counterparts, and each had a higher overall error rate, because they gave confident wrong answers where the free versions declined. Premium tiers buy speed and features. They do not buy accuracy.

Robots.txt did not reliably hold. The researchers found tools retrieving content from publishers that had blocked their crawlers. Perplexity’s free version correctly identified all ten excerpts from paywalled National Geographic articles, despite National Geographic disallowing Perplexity’s crawler. If you manage crawler directives for clients, that gap between stated policy and observed behavior is worth knowing.

One caveat carries real weight. This study ran in February 2025. Every tool in it has been through multiple model generations since, so treat the ranking as directional rather than current. The durable lesson is the one that survives any version change: these tools are confidently wrong often enough that your verification step is not optional.

How LLM Data Handling Should Change Your SEO Process

On February 10, 2026, Judge Jed Rakoff of the Southern District of New York ruled in United States v. Heppner that documents a defendant generated using a consumer-grade AI assistant were protected by neither attorney-client privilege nor the work product doctrine. A written opinion followed on February 17. Orrick and Harris Beach Murtha both cover the holding.

Read the sober version. McDermott’s analysis argues the ruling changes no privilege doctrine at all: it applies settled rules about third-party disclosure to a chatbot, and would have come out the same way had the defendant used a search engine or a library. The court also noted he was not seeking legal advice and was not acting at counsel’s direction.

The operational takeaway holds regardless. Consumer tiers are not a confidential channel for client information, and a personal paid subscription does not change that. Enterprise and API tiers across the major vendors contractually exclude your inputs from model training, and zero-retention arrangements are available on qualifying contracts. Consumer tier defaults are reported inconsistently and have changed more than once, so check the vendor’s own policy page rather than a summary article.

If your team expenses personal accounts for client work, you are paying for exposure. If you handle regulated data in legal, healthcare, or financial services, sort your contracts before you sort your prompts.

Frequently Asked Questions About LLMs for SEO

Which LLM is best for SEO? For most SEO work the four leading models perform comparably. Use Perplexity for client research, where citation accuracy matters and it has the best independent record. Use Perplexity or Gemini for GEO diagnosis, where you are observing citation behavior rather than relying on it. Use any current frontier model for strategy and copywriting. For technical SEO, use ChatGPT or Claude, which are close enough to each other that your verification process matters more than the choice.

Can AI replace an SEO? Not on current evidence. Keyword.com’s 2026 survey found 87% of SEO teams use AI regularly, but only 1% describe their work as fully automated. The tasks practitioners refuse to automate are the ones where errors are expensive.

Should I optimize for ChatGPT or Google? Google first, by a wide margin, then ChatGPT for referral traffic. Google reports AI Overviews at 2.5 billion monthly users and AI Mode at over 1 billion, while Conductor puts all AI referral traffic combined at about 1.08% of website traffic, 87.4% of it from ChatGPT.

Does schema markup improve AI citations? There is no strong evidence that it lifts citations on pages AI already cites. Ahrefs tested 1,885 pages that added JSON-LD and found no meaningful uplift on any platform. Keep schema for rich results and machine parsing.

Is it safe to put client data into ChatGPT or Claude? Not on consumer tiers, including personal paid plans. Enterprise and API tiers contractually exclude your inputs from training and support zero-retention terms. United States v. Heppner (S.D.N.Y., February 2026) confirmed that consumer AI conversations carry no privilege protection.

A Note on Current LLM Versions

Specific model names change monthly, which is part of why this article avoids them. The recommendations above rest on capability patterns that hold across versions: whether a model retrieves live, whether it executes code, and what the vendor’s data terms say. Check current flagship versions and pricing on each vendor’s documentation before acting on version-level detail.

Sources

Instead of scanning multiple search results, users now ask one question and receive a curated, confident answer. That shift is beginning to transform how people interact with websites.

These tools are cutting out much of the traditional top-of-funnel content. In some cases, users skip the research phase entirely and jump straight to conversion.

Real-World Example: AI Referrals Driving Direct Conversions

We built a custom dashboard for a client shown below, a prominent whistleblower law firm, to track referral traffic from AI tools. The results surprised us.

The top three most-clicked pages from AI referrals were:

This is unusual. In typical search journeys, users land on blog posts or service pages first, then move toward conversion. With AI referrals, users seem to be asking tools something like:

“Who can I report corporate fraud to confidentially?”

Instead of browsing, they are being sent directly to the firm’s most high-intent pages. Even more telling, form submission volume increased year-over-year.

Client Snapshot

Implications for User Conversion Funnels

This shift suggests the informational stage is happening inside AI tools. By the time users land on a site, they are ready to act.

Key takeaways:

Why SEOs Must Adapt to AI-Driven Search Behavior

Search journeys are fragmenting. Some still begin in Google, others with “Hey ChatGPT…”

Traditional SEO signals like backlinks and keywords remain important, but AI tools select results differently. They prioritize semantic relevance, topical authority, and structured information.

If your content is not being cited as a trusted resource, you miss out on high-value clicks. Worse, if you are not tracking AI referrals, you may not even realize your funnel is shifting.

A recent Ahrefs study showed that Google’s AI Overviews reduce clicks to websites by 34.5 percent. The behavior is similar across AI assistants: they summarize content and direct users to fewer but more trusted sources.

How Go Fish Digital Tracks and Measures AI Referral Traffic

At Go Fish Digital, our SEO team uses custom dashboards to monitor referral traffic from domains such as:

These dashboards capture both traffic sources and engagement behavior. We can see whether AI-referred users are more likely to convert, and often they are.

If you see high bounce rates on informational content but strong conversion spikes on forms or homepages, AI referrals may already be reshaping your funnel.

What’s Next: 5 Steps to Optimize for AI Referrals

1. Add FAQs and Q&A Content to High-Value Pages

AI platforms are designed to answer specific questions. Adding concise, direct Q&A content to product, service, and industry pages makes it easier for these tools to identify relevant information and use your website as a supporting source.

Focus on questions that reflect how customers naturally search and speak, especially questions related to costs, qualifications, processes, comparisons, eligibility, and next steps.

For example, a law firm might answer:

Place the most important answer at the beginning of each response, then provide supporting context. This structure serves users who want a quick answer while still giving search engines and AI systems enough detail to understand the topic.

2. Use Structured Data Where It Accurately Describes the Page

Structured data helps search engines interpret the purpose, content, and entities associated with a page. Where appropriate, implement relevant Schema.org markup for organizations, legal services, products, articles, authors, breadcrumbs, and other supported content types.

Schema should reinforce information that is already visible on the page. It should not be used to add claims, services, reviews, or details that users cannot see in the page content.

Although structured data does not guarantee inclusion in an AI-generated answer, it can make important information easier for search systems to classify. Coordinate with your development or SEO team to validate markup and monitor implementation errors.

3. Strengthen Your Highest-Intent Pages

AI-referred visitors may arrive later in the decision process than traditional organic search visitors. That means your homepage, service pages, contact forms, consultation pages, and submission pages need to answer critical questions immediately.

Review these pages for:

Avoid relying on vague marketing language. Clearly state what your organization does, who qualifies for the service, what users should expect, and what action they should take next.

For sensitive industries such as legal services, healthcare, or financial services, reduce uncertainty around privacy, response times, and the information required to get started.

4. Track AI Referral Traffic and Conversion Behavior

Create a dedicated reporting view for traffic from known AI platforms. This may include referral domains associated with ChatGPT, Perplexity, Gemini, Copilot, Claude, You.com, and other emerging tools.

In GA4 or a reporting platform such as Looker Studio, evaluate more than session volume. Track:

Compare AI referral behavior with organic search, paid search, direct, and other referral sources. A small number of AI-referred sessions may still be meaningful when those visitors are disproportionately likely to complete high-value actions.

Referral data will not capture every interaction with an AI platform. Some traffic may appear as direct, unassigned, or another source. Treat tracked referrals as a useful directional signal rather than a complete measurement of AI influence.

5. Monitor How AI Tools Describe and Recommend Your Brand

Referral reporting shows when users click through from an AI platform, but it does not show the full context of the recommendation. Regularly test the questions your prospective customers are likely to ask and review which brands, pages, and sources appear in the responses.

Look for patterns such as:

Use these findings to guide content improvements. If AI tools regularly cite competitors for a specific question, review whether your website provides a clearer, more authoritative answer. If your brand is described inaccurately, strengthen the relevant service pages, organization details, author information, and third-party profiles.

The goal is not to optimize for a single prompt. It is to build a consistent body of accurate, useful, and trustworthy information that AI platforms can retrieve across a range of related questions.

Final Thoughts

AI referrals are not just noise. They are a new form of high-intent traffic that bypasses traditional funnels.

If your site is not optimized for AI visibility, you risk missing some of the most valuable clicks. If you are not tracking these referrals, you may be blind to major shifts in your user funnel.

Start by measuring AI referrals, strengthening your most important landing pages, and reviewing how consistently your brand appears across AI-generated results.

Want to see how visible your brand is in AI search? Request a free GEO scorecard. 

What’s happening on TikTok right now is a compression of the entire purchase funnel into a single feed session:

The average time from scroll to purchase on TikTok Shop is under four minutes.  On traditional ecommerce, that same journey takes eight to twelve.

That gap, when compounded across thousands of purchases, is making a significant impact, and the brands paying attention to it are building serious revenue.

TikTok Shop reached $33 billion in global GMV in 2024. U.S. sales surged 120% year over year in the first half of 2025, and the number of larger brands with at least $30 million in annual revenue grew 95% during the same period. Those metrics demonstrate the dinosaur-sized footprint of a platform that has moved from cultural curiosity to legitimate sales infrastructure in under two years.

Key Takeaways

Why TikTok Shop Works Differently Than Every Other Commerce Channel

Every commerce channel that came before TikTok Shop required the customer to bring intent to the transaction. You searched because you wanted something. You visited a product page because you were already considering it. The ad for a competing product interrupted your attention and hoped to redirect it.

TikTok Shop generates purchase intent inside the platform. The conversion rate on TikTok Shop sits at 4.7%, compared to 1.9% across other social platforms. That gap exists because the content that’s doing the selling, a creator demonstrating, reviewing, or reacting to a product in real time, is indistinguishable from the content the user came to watch. There’s no interruption to overcome. The product is already inside the entertainment.

This is why the category breakdown so clearly mirrors the leading content types natively on the platform. Beauty and personal care dominates, generating over $1 billion in U.S. revenue, with womenswear following at $621 million. 

The Creator Model: Why Volume and Fit Beat Follower Count

The dominant creator strategy on TikTok Shop is not influencer marketing in the traditional sense but affiliate marketing at platform scale. And the math is a bit different. TikTok influencers with up to 50,000 followers have a 30.1% engagement rate for affiliate links, 1,57% higher than comparable Instagram influencers.

This is the central counterintuitive insight of the platform: the creators moving the most product are frequently mid-tier voices with highly engaged niche audiences, not celebrities with broad but diluted reach. A creator who has built trust with 25,000 people sharing a specific health concern, aesthetic, or interest converts at rates that no paid placement can match, because the recommendation lands with the authority of a peer, not the visibility of a sponsor.

The most successful brands on TikTok Shop have internalized this and built accordingly. Medicube, the top-selling merchant on U.S. TikTok Shop in 2025, generated nearly $100 million in revenue through a disciplined, platform-native growth strategy that leaned into creator partnerships, product innovation, and TikTok-native commerce moments. The brand scaled by identifying creators whose audiences mapped to their customer profile, seeding product consistently, and letting the content economics do the rest.

The affiliate model also changes how brands think about live selling. Rather than requiring the brand to host and manage live sessions itself, the live affiliate approach puts creators in front of their own audiences, featuring products naturally in sessions they already run. The brand scales reach without scaling operational complexity. The creator earns on conversion. TikTok’s algorithm rewards live activity with greater organic distribution, so the incentives stack in the same direction.

What the Data Says About What Actually Converts

Three product and content characteristics define the brands generating the most significant TikTok Shop revenue, and they show up consistently across categories and brand sizes.

Demonstrability. Products that can show (not tell) their value through transformation, mechanism, or before-and-after videos outperform those that require the viewer to imagine the benefit or extrapolate its relevance. This applies to beauty, fitness, food, apparel, and home goods alike. 

Price point. The $20 to $50 range represents 58% of all TikTok Shop transactions. Below that, the margin story gets difficult. Above $50, consideration lengthens and impulse behavior fades. The sweet spot is real, and brands that have optimized their TikTok Shop catalog toward it report meaningful conversion improvements.

Content volume over content polish. The platform’s algorithm optimizes for recency and engagement, which means a steady stream of authentic creator content consistently outperforms periodic high-production placements. Maëlys, the body-care brand that won the 2025 Glossy Pop Award for Best Use of TikTok Shop, became the top-selling body-care brand on the platform by anchoring its strategy in authentic before-and-after transformation content backed by a creator incentive program with prizes up to $275,000 for hitting GMV thresholds. The content looked organic because most of it was. 

How to Know if Live Selling is Right for Your Brand

No honest analysis of TikTok Shop can skip the live selling conversation, because the platform has pushed it heavily and the U.S. market has responded differently than anticipated.

Livestreaming has never accounted for more than 30% of U.S. TikTok Shop revenue share. Pre-recorded creator videos have steadily grown to represent two-thirds of TikTok Shop’s U.S. revenue. The live shopping behavior that dominates

TikTok in Southeast Asia reflects cultural familiarity with the format that the U.S. market hasn’t yet developed at scale. That doesn’t mean live selling doesn’t work here. There are instances where it can be very successful, but it isn’t the primary driver of conversion for most brands right now. 

The live affiliate model described above threads this needle usefully.

Rather than requiring a brand to build a live-first strategy, it allows live selling to develop organically through creators who are already comfortable in front of a camera and already have audiences who show up for them. The brand benefits from the format without carrying its operational weight. It’s like finding Goldilock’s lovely happy medium.

What to Actually Do if You’re Considering TikTok Shop

What Go Fish Clients Have Built on TikTok Shop

At Go Fish Digital, we specialize in helping ecommerce brands seize the platform by the algorithm and make the most out of their TikTok Shop presence.

What connects these results is the same thing that separates TikTok Shop winners from TikTok Shop experimenters in general: the brands that scale have systems. Systems for creator identification, onboarding, and content guidance. Systems for measuring what’s working at the product and creator level. Systems for connecting TikTok performance to the broader revenue picture.

Ready to find out if TikTok Shop is the right channel for your brand? Request a free social commerce audit and we’ll tell you exactly where you stand.