Every business—but DTC brands especially—have to weigh opportunity costs. In smaller brands, for instance, it’s fairly common to focus exclusively on acquisition. After all, there are so few customers to retain that any business value you might gain from that is going to be marginal. Spending time on retention could, in fact, seem wasteful, because it takes away from time that could be spent on something higher value.

You could see this calculus changing, though, when AI can think about—and do—the work for you.

Even if those early retention gains are marginal, if you don’t have to do the work (much less think about the work), why wouldn’t you take them? It would almost feel wasteful not to.

This is the piece I think we’re missing: AI will very often flip on its head the calculus of whether work should get done. If that’s the case, more work will get done.

This, then, poses the question: Which work should AI do?

There are a few potential answers:

• AI does the work that delivers the highest confirmed ROI, because it is obvious that work needs to get done.

• AI does the work that delivers the highest speculative ROI, because the business agrees it is work that should be done—but it doesn’t have confirmation that the work will deliver value.

• AI does the work that delivers low, but confirmed ROI, because this is work that is not currently being done and—when added up—will deliver meaningful incremental gains to the business.

Sitting here today, I don’t think you could say any of these choices are bad or wrong. The choice is probably more a reflection of how much you trust AI to do the work today than which work you want to do.

Because, in reality, if you trust AI to do some work, you’ll start doing all of this work—which is different from the way most businesses operate today. That’s a big change.

But the specifics around execution will be interesting to watch.

But when all that talk about change starts to get tossed around as the big trend to watch for the upcoming year, I end up reminded of the Jeff Bezo’s quote that inverts the question around change. It goes like this:

“I very frequently get the question: ‘What’s going to change in the next 10 years?’ And that is a very interesting question; it’s a very common one. I almost never get the question: ‘What’s not going to change in the next 10 years?’ And I submit to you that that second question is actually the more important of the two—because you can build a business strategy around the things that are stable in time.”

And, so, this is where my predictions, if you will, come into play.

There is a ton of change in ecommerce. It feels like bigger shifts are happening than anyone within the space can even wrap their heads around. Of that, we can all agree. But it is unlikely that a prediction in the AI space—be it for internal operations or as a consumer channel—is actually actionable right now (to Bezos’ point).

What is actionable, though, is what is unlikely to change. And one of the things that feels unlikely to change is the importance of listening to what your customer has to say.

This is, of course, a bit self-serving and biased. But, then again, the reason I am here as the CEO of Stamped, KNO and Repeat is that I fundamentally believe in the permanence of listening to your customer.

You may read predictions about AI models cutting photoshoot costs or AI models replacing UGC/traditional ad creative. You may read about ChatGPT becoming the channel to figure out. All of this—and more—could be true.

But what would your models look like and what should your ads say and how will ChatGPT know whether to recommend you? What your customer says about your brand and your product is the input to all of that.

At the very least, you need to know what they’re saying; at the most, you need to take action on it.

And that prediction is certain to come to fruition.

As the dust settles, though, one thing is clear: Everyone wanted to participate in the AI story.

Adobe, which saw $11.8B in transactions on Black Friday alone, said AI-driven traffic was up 805% year-over-year. Salesforce, which reported servicing more than $336B from November 25 to December 1, said agentic customer service conversations increased by 55% over the previous week. And Shopify, which was quiet on sharing how consumer AI impacted the $14.6B its merchants did, came out this week and just hammered home the AI narrative at Editions.

And the AI story isn’t just that it exists. Adobe said visitors who came from an AI chatbot were 38% more likely to make a purchase.

As someone who has said that it might take longer for ChatGPT to make its way into consumer’s browsing and buying habits, I find this datapoint interesting.

• Are we adopting way faster than expected?

• Is 805% growth more a function of hardly any usage last year?

• Will this persist?

The news on AI, especially from a consumer market perspective, seems to constantly shift. OpenAI has declared a “code red” to improve ChatGPT, because Google’s Gemini is challenging ChatGPT’s position in the market.

If the presumed winner actually loses—or just doesn’t win by as wide a margin—what about this year’s AI <> BFCM narrative actually sticks?

Maybe there’s a lesson here.

As the tech grounds continue to shift below our feet, maybe the best approach is to skip trying to find the right attachment angle and continue focusing on the core of what makes commerce go: People buy products (and into brands), they talk about them, and, as they do, more people buy. Everything else is just way for those things to happen.

In looking back on last week and weekend, it does feel like a unique time where every part of a brand has to work together in a more intimate fashion: Marketing can’t win alone. Ops can’t save a bad promise. CX can’t polish a broken experience. It all has to fit together, and that usually happens through great planning.

One of the places, I think, that’s often overlooked for qualitative inputs to that planning is reviews, and I think that can actually pull double duty during holiday—serving both as the “social proof widget” that boosts conversions and as a data stream that basically becomes planning infrastructure.

Hear me out:

Reviews are the one artifact everyone touches: They reveal what customers expected, what they actually experienced, and how they felt about the trade-off. Unlike NPS or CSAT, they aren’t designed around an internal agenda, and the free-form nature means customers choose the topics. They’re also timestamped to a specific moment, so if you can cross-reference them with what you were doing internally at that time, you get insights around operations, product quality, and whether a customer’s expectations were met based on the advertising they saw.

That makes reviews a shared truth object.

Marketing can take the phrases people use when they’re happy and build future positioning from the language that already converts. CX can focus on the issues that appear consistently — not just the loudest edge cases — and shape better support content before the next busy cycle. Ops can look at shipping, packaging, and fulfillment misses that cut across SKUs and turn that into a prioritized list everyone recognizes as real. Product and merchandising can zero in on the “almost perfect but…” patterns that point directly to fixes with immediate revenue impact.

And, interestingly, reviews you collect from BFCM purchasers might be even more valuable, because they are—contextually—from a time when all your internal teams are the most dialed in. So, all those reviews are from when you’re performing at your best.

The opportunity, then, is to use them as inputs to planning so you can close that gap between expectation and reality—something that’s probably even more valuable than displaying them on a PDP.

One of the pieces we’ve failed to talk about, though, is how the end user begins to shift as you begin utilizing reviews in those different fashions. While consumers are the primary beneficiary of traditional PDP-located reviews, that beneficiary label ends up getting shared with the brand/marketer as you expand the surface area for reviews. And that’s mostly because the application is different.

While reviews in this “new structure” are still content getting served to consumers, they are not pure-play reviews; they become inputs to the content and those inputs need to be synthesized by the marketer/brand. This, as we’ve discussed, is already happening with ChatGPT and some marketers, but we believe that’s going to continue changing and adoption will accelerate. And, when that happens, the primary benefit of a review isn’t so much helping a single customer as much as it’s creating business intelligence for a brand to help a multitude of customers by better positioning, better messaging and better product improvements/expansion.

When that happens, the primary beneficiary of reviews flips from the consumer to the marketer/brand, because the insights accumulated become inputs that make multiple surface areas of the business better, not just a purchase decision.

Because of the distributed nature of these benefits, the measurement of such implementations isn’t as straightforward as “customers who engaged with reviews versus customers who didn’t,” other proxies for review effectiveness, or even A/B testing whether to show reviews. What you’d expect, though, is that you’d end up with the right inputs to improve each surface area where the business intelligence is being applied. That might be ads, that might be website copy, that might be inspiration for product expansion or product improvement.

In each case, though, it’s the marketer/brand with more information to make better decisions. And that will make them the main user.

This is happening quickly, though, and the funnel is a pretty big, pretty nebulous thing. So, maybe a model for thinking about this is where parts of the funnel happen today. If you were to map to consumer interaction points, it probably looks (roughly) like this:

• Awareness

  • Advertising

  • Word of mouth (both IRL and online)

  • ChatGPT

  • Retail Distribution

• Consideration

  • Social media

    • Instagram

    • Facebook

    • TikTok

    • Etc

  • Research

    • ChatGPT

    • Reddit

    • Search

    • Word of mouth

  • Brand website

    • Product Listing Page

    • Product Detail Page

• Conversion

  • Brand website

  • Retail

  • ChatGPT (not yet, but soon?)

Reviews, today, sit in basically two places:

1. As close to the point of conversion as possible without interfering with it (the product detail page)

2. All the way at the top of the funnel in awareness (word of mouth is pretty much a review; if you tell someone about something you like, you’re both introducing them to the product and offering a positive review of it)

Everything else? They’re fairly open spaces for reviews.

As an example:

Brands, though, have leaned into user-generated content in nearly all those spaces, encouraging and rewarding photosharing on Instagram, partnering with influencers to push their products in videos, and paying affiliate fees to blogs for steering traffic their way. If you squint, you can see a review inside all of these.

As brands begin shifting some of their strategies from ranking highly on Google to showing up more frequently in LLM answers, though, photo dumps and “empty” endorsements won’t be enough to feed the content beast of ChatGPT. It will need more context.

Brands, then, will look for those customers and creators to provide it.

Customers, of course, will handle that on their own. They already do. But creators will need to know what the overlap is between what their audience likes and what a brand’s customers like, so that the context can be specific and helpful. Where might you get that information?

The obvious—and most robust source of this data—is your reviews.

Today, it lives in two places. Soon, it (and by-products from analysis of it) will live everywhere.

Most of those explorations, if you will, were mostly about how they are getting used and will get used in the future. (We’ve talked about how reviews appear to be informing LLM recommendations, how reviews might be pulled further up the funnel during “value-led buying” time periods, and that reviews (along with other customer data) will be used to actually distinguish otherwise identical ads in the future (since creating them will cost next to nothing).

What we haven’t talked as much about, however, is how early adopters are doing this right now. There was a post on LinkedIn that gave a very practical—albeit brief—guide on how to do it:

These instructions are pretty easy to follow, and that’s one of the benefits of agentic workflows: So long as you know what prompt to use, you can get to what you’re looking for a lot faster.

Community sort of solves the first part (“what prompt to use”), because people will share what works for the task, but the second part is actually harder (“what you’re looking for”). Yes, you can learn what work to do and how to do it, but you still need to have an understanding of what you’re looking for and why you’re looking for it.

What I love about this post is how illustrative it is in terms of how much faster we’ll begin to copy what’s working, but also how much it removes the need for critical thinking from the marketer.

Copying this prompt (and others like it) might feel like a win to be moving faster, but I think that will ultimately slow marketers down in the long run when they run out of things to copy.

The real win from AI won’t be getting you the information; it’ll be helping you understand why you need to be looking for certain information in the first place. And, eventually, I’d bet it’ll just do it for you—no copying required.

Even I am guilty of this, and one of the “indicators” I look at around this time is the answer to the question: “Do you and your family plan to purchase more or less online in the next 3 months than normal?”

Every year, October sees the biggest spike. And it makes sense: the holidays are finally in sight for families.

This year, though, our numbers are lagging recent historicals (2023 and 2024):

I don’t know what to fully make of this.

I’m sure part of it is inflationary/macro-economic in nature (if everything costs more, I can’t spend as much on holiday gifts as I normally do), but I do wonder if some of this is also cultural: We spend so much time online now, and we’re beginning to spend so much time with AI… is there a desire to spend some money offline?

The numbers, if you look closely enough, have been trailing historically since December 2024—almost a whole year now. Oddly enough, this started when inflation started to calm down, so it feels like it couldn’t just be that. I don’t know?

But, even still, the cause might not be as important as the trend, especially if it means some amount of dollars are shifting at all.

For brands with retail distribution or their own brick-and-mortar operations, it could be a question worth asking, because the answers may inform future execution. Where should you be? Which channels should you be investing in more right now? Does discovery in one channel lead to a transaction in another? What sort of word-of-mouth and social proof is required to optimize each channel?

I don’t know the answers yet, and I don’t know if it’s even right, but it’s something worth paying attention to. Even if it’s just to look ahead.

There are trade publications that write about this, there are industry voices who research it, and there are software companies that plot trends around it.

Now, there are even academic papers on how well those citations and recommendations convert.

I find this last piece interesting, because it suggests that LLM’s perform poorly when viewed through the lens of last-click attribution.

But you know what else performs poorly when viewed through such a lens? Meta advertising.

Discussions around LLMs, it seems, have jumped very far ahead in terms of compressing the funnel. ChatGPT’s announcement with Shopify, Etsy and Walmart are all exciting, but, right now, it looks like ChatGPT is primarily a discovery channel.

And the way discovery works is pretty universal: People talk about things they like. The more they talk about them (and the more people who talk about the same things), the more other people hear about them. Then, the businesses that sell those things grow.

This is user-generated content at its core. Reviews are the simplest distillation of that. And we traditionally have thought about them as a bottom-of-funnel mechanism that helps convert.

But what if we’ve had that wrong? What if reviews are only a bottom-of-funnel mechanism that helps convert, because that’s where we’ve primarily used them?

At Stamped, we know plenty of marketers who read through reviews and find the best ones to use in advertising, email, etc. They’re effectively bringing reviews up the funnel, sometimes even to the point of discovery.

And if we end up in a world where LLMs enable the same amount of discovery as Facebook and Instagram, it’s going to happen, in part, because there’s an incredible amount of people who talk about the things they like. And that will make reviews more important than they ever were—because their impact will be seen throughout the entire funnel.

There was one comment in there that was so disarming in its honesty, and it’s worth just showing the screenshot over trying to summarize it:

My first reaction wasn’t “that’s ridiculous;” it was: “Is MacCoy right?”

The Shopify ecosystem, at this point, is pretty mature in terms of the technology — at least through the eyes of the brand. So mature, in fact, that a healthily 8-figure brand thinks returns software should cost about $25/mo, since they know you can pay $15/mo for reviews.

Email is email. Not much innovation there at this point. SMS is a newer channel, but the feature set really just needed to hit parity with email to appease power users. That’s largely happened. Returns softwares mostly do the same things now. Reviews are table stakes for any brand. The list goes on.

And that’s sort of the point: Every category has reached a level of “good enough” that any improvements make things marginally better than they were before, but not meaningfully better. If I’m a brand, why is any of it worth a premium now?

This is the challenge in any space (not just software): Something new comes along and it’s worth a premium. Eventually, others copy it and it’s worth less—because both competition and time turn a novelty into a commodity. Once the expectations around value are met by multiple players in the same category, the value ends up getting “priced in” to the cost.

And because those premiums are eroded, the market comes to expect this cycle, meaning they just know that the new vendor in the new space is overcharging. Time will prove them right. Just watch!

Vendors, then, have two options:

• Get efficient enough to reduce prices

• Innovate enough to justify premium prices

There is, though, another way: Innovate and keep prices lower than premium. At Stamped, we think there’s an ability to do that in reviews. Word of Mouth is only increasing in importance, and the fundamental components of that are best seen in reviews.

None of this is that easy to do. But there is a lot of truth in that Twitter exchange, and it’s a net positive for everyone.

And while I don’t want that to happen, I do want to spend a minute on Sora, the new app that’s growing faster than ChatGPT even did. But it’s probably not for the reason you think.

Sora, if you’re not familiar, is an AI video generation app that’s got social network components to it. It’s like TikTok + Vine + ChatGPT rolled into one. And it hit 1M app downloads in less than five days while being gated by an invite flow that required prospective users to get a code from a friend.

The content that’s coming out of Sora, though, is something brands need to be paying attention to.

For years, brands have moderated user-generated content, showing only the best reviews onsite, using only the best photos from customers in their ads and instagram feeds, and replying to negative reviews in public places with “let’s make it right” type messaging. Many of us in the commerce space have talked about how consumers have more power than ever, but, when it comes to using their voice, they still play by the brands’ rules. UGC, for all its authenticity, ends up being a manicured look.

Sora content, which is quickly becoming popular content in more traditional channels like Twitter, Instagram, etc, is unhinged by comparison. There’s a video of Mr. Rogers arm-wrestling Bob Ross. And a video of a Michelob Ultra can getting arrested for public indecency. And another of people making their friends cause a scene in a KFC.

I can’t imagine PBS, AB InBev or YUM! Brands are thrilled about this content. And that’s exactly the point.

If “AI slop” becomes a popular form of communication (and why wouldn’t it? Memes are already the language of the internet), then it’s pretty clear from the earliest creative outputs of Sora that consumers now get to talk about brands in the way they want to, wherever they want to.

I don’t think there’s a way to have a full appreciation for what this means, except to say that this will be the biggest shift in user-generated content we’ve ever seen. Consumers aren’t creating content for your products anymore; your products are now for their content.

You can bet it will be messy and weird as we figure it out, but you can also bet that it’s going to happen. The best you can do right now is understand what your customer thinks of your brand and your products.

There was a lot of talk this week about the big OpenAI/Shopify partnership on Agentic Commerce. If you missed it, well … you can read about it here, here, here and here. And here. (And a bunch of other places. It’s a big deal.)

The thing that’s gone underdiscussed (or maybe even not discussed) is the role of reviews. So, we’re going to cover that.

First off, if you just look at the experience, it’s pretty clear that content matters a lot.

Here’s a video of a fairly general search for dinnerware. Take a look at what ends up showing up in the results:

I’m not talking about the brands. I’m talking about the connection between the content (pulled from context … market trends in this case, since I was using a temporary chat) and recommendations. And, then, look at the connection between the supporting content on the recommendations (“Why you might like this” and “What people are saying”).

There’s a through line all the way from the context to the supporting content, which makes it appear that reviews are powering a good deal of what gets recommended to a person.

For a while now, reviews—which everyone agrees are important—have been a category that’s been favored for commoditization (how do we make this thing cheaper?) over innovation (how do we make this thing more valuable). I think this is likely the signal that reviews are back.

And I’m super excited for us to put our spin on it.

While I disagree with this in practice, I love the spirit of it and I actually think it highlights an opportunity for SaaS: Empathy. In short, there’s too little of it—and if you have it, you can overcome not running a DTC brand.

In fact, it might be even better.

Running a brand means you have a single point of view on solving a problem. Running a SaaS company means you get presented with dozens of points of view on solving a problem any time you talk to customers. Which one is best? Which one works for the most number of brands? Which one can be modified to work with even more?

These are the types of questions SaaS companies get to ask and answer, and doing so can create a much better solution than the one a brand might come up with on its own.

It is, really, why ecommerce-enablement SaaS became a market after all: Brands have a core competency, and it is often not building software. They know their problems intimately, and they can often solve them (DTC brands are scrappy!). But they can’t solve all of them, because there’s an opportunity cost to doing so (the more they solve their problems, the less they focus on their product). SaaS vendors, on the other hand, have a complementary core competency, and it’s being adept at solving problems.

But that only happens when a SaaS vendor is willing to invest in understanding its customers.

As generative AI gets infused into more and more workflows, that empathy is going to matter a lot more. A strong point-of-view on how to do something is the best way to improve prompts, for example, and it will matter much more when it comes to automated workflows.

Those who take the time to understand will be the ones who deliver the most value.

You don’t need to run a brand for that. You actually have to do a lot more. And I’m betting on Stamped being one of those vendors.

For starters:

• Google announced AP2, an open protocol for payments to be handled by agents. (Google said they’ve been working on it with more than 60 organizations, including Adobe, Salesforce, American Express and Mastercard.)

• Shipt announced a partnership with Perplexity at Shoptalk in Chicago, that will let people build carts in Perplexity’s Comet browser off what you might call otherwise use as a search query for inspiration. (It’s shrinking the discovery > conversion funnel path to a single “search”)

First off, each of these are cool developments on their own.

But haven’t we been trying to collapse the funnel for as long as we’ve been selling things? Consumers are human and humans need time.

It might be cool (heck, it is cool) to think that, at some point in the future, you could type “kids bedroom remodel ideas” into an AI browser, flip through a bunch of photo-realistic mocks of your daughter’s room and click a button to have all the decor from that very “photo”—down to the right amount of paint—shipped to your house.

It also might be completely anxiety-inducing if you’re a brand. How do you make sure your brand is in that remodel idea? How do you make sure a customer thinks about your brand when they see those remodel ideas? How do you actually plant the seed that a customer should remodel their kid’s bedroom?

And these are reasonable questions to ask. I bet some consumers would do something like this sometimes. And, in those circumstances, you might miss a sale!

This is the story a lot of people are painting: That agentic commerce will erode brands to inputs in a generative AI model and any semblance of brand-building is going to be moot.

The problem here is that the pervasive version of the future flips to everyone doing this all the time. That feels wrong.

And while there are a lot of reasons I think that’s wrong, the biggest one is that, for agentic commerce to become the primary consumer behavior, it will require all brands to be competed away to commodities.

If the history of commerce has taught us anything, it’s that the more media space there is, the more brands there are. And that’s because, the more money consumers have (which allows for more media), the more interests vary.

I may value a specific product category far more than you do: You may treat that category as a commodity. But me? I care so deeply about that product category that I’m willing to pay significantly more for that category than you are.

And you know what? The world is big enough and economies are strong enough that there are a lot of people in both camps.

Maybe agentic commerce means the bottom of the market gets competed away. But there’s always going to be room for brands.

The old Shopify mission of “arming the rebels” is probably a little tired now. And maybe, actually, it’s complete.

Shopify used to say this all the time as a way to position itself—through the brands that used them—against Amazon (or big retail more generally). Part of that was to give them the types of tools (and the quality of tools) that Amazon had built for itself; the types of tools that create customer experiences that, so long as the price is fair and the product is good, encourages a customer to come back.

Those tools pretty much exist!

In fact, one of our latest reports—Fractures in the Funnel—includes data that shows customers now have increasingly positive sentiment toward buying directly from brands and increasingly negative sentiment toward buying from marketplaces (like Amazon).

That’s a big win. But it isn’t a complete win.

As sentiment shifts, so, too, do expectations. That means continuing to deliver great customer experiences is now required to keep sentiment high. And while the tools are there, they’re not always being used—or not always being used in the way that can elevate customer experiences even more.

How do you solve that?

I’ve talked a lot before about how much room we’ve had at KNO to improve the way we speak about our product, promote features and explain their value in hopes of getting customers to use more of it. That responsibility is on us as a SaaS vendor more than it’s on a brand.

But it’s also an opportunity for brands to press their SaaS vendors.

One of the biggest hacks that exists for brands is to ask their SaaS vendors what they’re seeing from their other customers and what they’re seeing in their data. They often have far, far larger datasets than brands (or the accumulated datasets that a brand can gather from peer networks), and that volume of data can help keep brands ahead of consumer expectations.

It’s a cool way to partner, and I’d encourage you to think about it. And, if you want to press us, please do!

Sounds funny, but hear me out: Every month, we look at our consumer confidence index as part of the DTCCI dataset. And while it’s not down over last year (a not bad sign), it is down over August and the year, overall, has been trailing 2024. I think we’d qualify that as “not good.”

But what happens when confidence remains suppressed for a prolonged period of time? Do people still review products in the same way? Or do people’s expectations change? And do they express frustrations more readily?

These are questions I think are worth exploring, and I think this type of contextual awareness (much like the consumer confidence index we share) could be valuable to reviews strategies and loyalty.

You might want to change review moderation or wait longer to ask for reviews. You may want to prioritize reviews that include higher quality photos to better engage people evaluating your product for the first time and reinforce quality and value.

The list of tactics that might change is fairly comprehensive and, also, fairly simple to execute on. And it just might be that it’s worth keeping a pulse on what the macro environment is like to decide when to change those tactics.

He was talking about AI and its impact on software, basically saying that a software company’s product is fairly indefensible now. The speed to reacting to a competitor’s new feature has been basically reduced to zero, so the last-standing piece of defensibility is brand.

Jason said that, roughly speaking, 95% of people who are good candidates to buy from a category are going to end up choosing a product based on brand. Interestingly, though, he feels everyone is back in the market right now. (Maybe for the first time ever?) AI is that disruptive.

My first thought here was … wow, that’s basically consumer. You never really own a customer and they are always in the market, even if they continue buying from you.

But the other thought, and the one I think worth spending a minute on, is how AI will both perpetuate that consumer reality and take it away.

In consumer, the number of brands a market can support is roughly equivalent to the distribution advantages created by media. Consider:

• Newspapers as the primary media source: few brands per category

• Radio as an emerging media source: more brands per category

• TV as the primary media source: dozens of brands per category

• Social as the primary media source: hundreds of brands per category

What happens when AI becomes the primary media source for consumers?

Pre-social/pre-algorithm, every consumer was aware of nearly all brands in a category, even if recall was just a few. Your goal as a consumer brand was to increase frequency such that you were recalled at the right moment.

That is impossible now. How do you beat the algorithm to do that?

The flip side of that, though, is that the algorithm allows for more brands to exist, because it segments the market for you. That may limit your upside, but it also increases the likelihood of some sort of success.

AI (at least where it appears to be going) sort of has the algorithm beat, in the same way the algorithm had TV beat. It knows (or will know) you incredibly well. Not just your content preferences, but your product preferences and aspects of your lifestyle that may create leading indicators around new product categories to explore.

All of that means room for more brands. Which means there will be more competition and a necessity to both more deeply understand why your customer buys your product specifically and how you can communicate that—both to the consumer and to the AI.

Markets, it seems, are going to continue getting bigger and continue getting more crowded. If that’s the case, the question will be how successful you can be in keeping your customer from going back to the market.

This month alone, there’s been coverage in places like the Wall Street Journal, Barron’s, New York Times, CNN, NBC News (the list goes on) have been talking about whether prices are rising because of tariffs and whether tariffs will have an impact on Black Friday/Cyber Monday.

Some retailers, like Walmart and Home Depot, have even said they’ve been absorbing tariff-related price increases, but will need to start asking consumers to do that instead. And Goldman Sachs published a report that said consumers will absorb 2/3 of tariff-related price increases by October.

That’s a lot of news in a really short period of time. Especially when it sort of felt like it was fading to the background.

While I’m not interested in arguing about whether all of this talk is right, it is interesting to me that, in the midst of all this talk, that DTCCI data has shown present spending from consumers plunging.

Are the two things related? I don’t know.

But it certainly feels like consumers are hearing that things are going to get more expensive for them. And that, I think, is really the most important part: the emotional and psychological effects of fear around rising prices. Those two things can change consumer behavior before anything else.

If I were running a brand, then, I’d really want to understand a good amount from my customer about where I fit in their lifestyle. Am I a luxury? Am I a candidate for trading down? Am I a better value buy in the future if my competitors are forced to raise their prices and I’m not?

There’s a bunch of other pieces, too, but if I can start to understand if I’m a value buy or a brand buy for my customers, I can do a lot to adjust my merchandising efforts and incorporate levers that protect my base and, hopefully, expand my base by finding the right levers to pull around value and brand.

And, like a change in consumer emotion might precede a need to actually change behavior, a brand’s ability to precede the change in consumer emotion better positions the brand to win when that happens. This feels like it would be a good time to do that.

One of the reasons, I suppose, to keep on this topic is that BFCM is a huge value-led buying moment. It’s become commonplace to discuss how much “worse” customers are when they’re acquired during BFCM, but, when I look at KNO data, I come to a different conclusion. Most customers acquired during BFCM sales are just driven by value. And they rarely see that type of value again.

Despite the cultural narrative around gift buying during BFCM, our KNO data from past years shows that nearly 3/4 of all purchases aren’t made for anyone else; they’re made for the person making the purchase.

Add to that the fact that people buying from a brand for the first time during BFCM often have known about the brand for much longer than normal, it seems highly likely that a good deal of customers acquired during BFCM are price-elastic customers. They’re more willing to buy when the price more heavily changes in their favor.

If that is, in fact, the case then it’s worth asking how many other times per year those customers can find a similarly good deal on your product.

The answer? It’s unlikely they can.

When we started looking into this concept of value-led buying after Prime Day, we looked at the average discount percentage by holiday. Nothing comes close to matching BFCM.

The question, then, is what to do about it.

If you’re willing to acquire a customer at a specific discount, are you willing to retain them at that same discount? Or would you prefer to only sell to them once?

The answer to this question likely adjusts your 1) retention strategy and 2) BFCM sales strategy. And with Labor Day right around the corner, you have an ability to test run any changes you want to make.

We’ve spent some time here over the last few weeks talking about value-led discovery and how UGC, social proof, and word of mouth can justify that value.

One of the pieces we haven’t spent as much time talking about: What happens after the initial purchase?

We’re going to be pulling data around this question, so this is more of an “experience of one” musing, but I’m curious of your experience, too. (So, request: put on your consumer hat and tell me whether this is relatable.) And the best place to start with this, I think, is that the price I pay for something for the first time I buy it is the price I expect to pay going forward.

So, for example, if I get 20% off a $50 product for providing my email address, my true cost on the product is $40. And I really won’t pay more than that again; you’ve told me that you’re willing to sell the product at that price and think that’s a fair value. If I think that’s a fair value, too, why pay more?

This, at least, is my personal (and even somewhat emotional) rationale, but I suspect others are like this, too.

And it’s what happens after this that I think is most interesting. Because what I’ve observed in my own behavior is that one of two things happen:

1. I find the product (either DTC or in retail) on sale at the price I’m willing to pay and stock up. But I only do this when I need the product.

2. I can’t find the product on sale at the price I’m willing to pay and buy a new, competing product in the price range I paid originally. Sometimes, I end up liking this product more (sometimes, too, this product is on sale).

What this ultimately means, then, is that I end up in a cycle where brands are competing to win my share of wallet and, when they do, they pull forward a number of sales (essentially, taking me out of the market for a while).

This is fairly category specific, and, in some cases, I don’t necessarily end up at option 2 above.

An example: There’s a granola my family likes that I buy at Costco, but it goes on sale occasionally and I know that price is meaningfully better. I also happen to have a sense of the sale cadence. So, when it is on sale, I’ve done the math in my head of how many bags I need to buy until the next sale.

My purchase frequency isn’t really that good, but the brand has a near complete share of wallet for the category and it constantly pulls forward my demand. That, though, comes at a tradeoff. They’re taking more raw profit dollars today (because of the bulk-buying from the sale) instead of less total profit dollars over time.

That seems to work for them, and it definitely works for me.

But other categories? If a product isn’t as important to my family for a particular reason, I end up at option 2, and whoever has the sale at the right time (or delivers the most value at the right time) ends up capturing my demand. You can see how that could be considered a lack of loyalty from a brand side, but, in reality, I’m fairly loyal to all of them. Just at a specific price.

These types of consumer tradeoffs, I think, are going to remain fairly important for us to consider as consumer sentiment remains low and value-led buying becomes a learned and, possibly, accelerated behavior with LLMs making deal-hunting much easier.

I’m curious what you think. Does this land?