S07, E14: Re-evaluating agentic commerce (with Andrew Lipsman)

Mobile game developers no longer need to pay up to 30% in major app store fees. With Xsolla Web Shop, you can create a direct storefront, [music] cut fees down to as low as 5%, and keep players engaged with bundles, rewards, [music] and analytics. Start today at xsolla.com. That's x [music] s o l l a.com, or use the link in the episode show notes. The problem is that the distinction needs to be drawn between [music] the competence of the economists and the correctness of their analysis. Welcome to the Mobile Dev Memo podcast. I'm your host Eric Seufert, and I'm joined today by Andrew Lipsman, who is rejoining the podcast for the third time. Andrew, welcome. Thanks, Eric. Great to be back. So, it's good to have you back. We spoke a little more than 6 months ago, and I wanted to kind of use this conversation today to take stock of

what's happened in the media landscape, the advertising landscape since then. But before we jump into that, can you please introduce yourself or reintroduce yourself to the audience? Sure. I'm Andrew Lipsman, an independent analyst and consultant. I write at the Substack Media, Ads, and Commerce. Previously, some folks may know me as a an eMarketer analyst, where I covered retail and e-commerce, and then started covering retail media specifically. So, a lot of people will know me more as a retail media expert. Yeah, and I'll link to the blog in the show notes, but yeah, I recommend that, you know, anybody who's following the space should follow your blog. It's great, and you provide a lot of really fantastic insights, particularly in the retail media space, but also broadly in the digital ad space. So, so like I said, we spoke just a bit over 6 months ago, and if I remember correctly, the sort of genesis of that conversation was that I'd written a piece, or I'd written on LinkedIn, and I was just kind of I was professing some skepticism towards agentic commerce as such, right? As it was being kind of depicted.

And then, you kind of jumped in on that. And I think we should sort of shared a similar sentiment. And then, I'd written a piece that that went I'd say it went fairly viral, called Agentic Commerce Is a Myth, right? And and the title there, I think, wasn't fully reflective of the point I was making in the piece. It wasn't that agentic commerce will never happen, or it's sort of like fundamentally impossible, anything like that. My my point was that it was the independent agentic commerce faced a lot of headwinds, faced a lot of frictions because of some natural tensions with the the big retail platforms, and also just what I thought was the superiority of the advertising model relative to that kind of affiliate links model that Open AI ended up introducing and then walking away from. So, so that's just to set the stage. But so, you know, we had shared an outlook at that time that agentic commerce seemed unlikely to take root. And that was a very unpopular position, it turned out, right? It Our position that we had shared at that 6 months ago podcast, that was very unpopular.

But now, here we are, almost May, and enthusiasm for agentic commerce does seem to have waned, especially now that ChatGPT's abandoned instant checkout, and Walmart, which was the largest retail platform to partner with ChatGPT, has integrated Sparky into the that platform as an app. So, they've sort of walked away from that sort of affiliate integration. So, what is the state of agentic commerce now? That's kind of I want you to set the stage for us here. What's the state now? Yeah. Well, let me rewind really quickly to that 6 months ago, because when you finally articulated the skepticism, I was just coming out of conference season, where I was talking to a lot of retailers, and they had become fully convinced in like a few weeks' time that all their traffic was going to go away in the next couple of years, that their retail media businesses were going to be, you know, vanquished, and that there was this crisis of epic proportions coming. And everybody seemed to buy into this hook, line, and sinker so quickly, and I was like, "But the consumer behavior has to happen first." That's my bias, my lens that I

always look at things through is consumer behavior. And I just didn't see it, cuz I'd seen so many trends, you know, where the technology always led the consumer behavior, and it just never manifested, voice commerce being kind of the most obvious one in this space. So, I voiced some skepticism as well. My My article is titled Agentic Commerce Is a Collective Hallucination, because I saw the industry collectively hallucinating this thing into existence that I didn't think was going to happen. And a lot of the conversation, you know, became a definitional debate what is commerce. And so, I think the definition that you and I used, and certainly that a lot of the pro-agentic people believe in, is this fully autonomous, end-to-end sort of decision-making on behalf of the bots, and bot-to-bot communication and transaction. And even if there's minimal human intervention, I just don't think this will ever happen. So, what's happened since then as we've had a series of things that have happened that have shown that the behavior really isn't taking off, and probably not likely to take off, is that we've moved Not we.

The pro-agentic crowd has moved the goalpost to include anything that is AI-assisted, or any transaction that might start in an LLM, or even on Amazon Rufus or Walmart Sparky, is now agentic commerce. I I don't totally buy it. I think we've all kind of started to use agentic commerce as a shorthand for all of this stuff, but I'm a proponent of like AI assistance. And then, even more recently, you know, so I think Open AI, obviously, the big news was that they, about 6 months into this instant checkout experiment, backed away from it. Walmart had said that, you know, it just wasn't converting, it wasn't working. No surprise here. And now, Open AI has sort of moved to a new strategy, which is the super app strategy. So, it wasn't working kind of plugging into the LLMs organically to have checkout there. Now, they're asking the retailers to take extra steps and actually build, you know, mini apps within the ChatGPT app, essentially. And this is another thing that's kind of a fever dream of all the big tech

companies over the years. They've always wanted to create the super app, and they emulate what's happened in China with WeChat. But that's another thing I'm very skeptical of. WeChat came out of a moment in time in which the majority of Chinese internet users weren't on the internet. And so, you had mobile, social, and e-commerce all happening at once. And as those things evolved together, you actually had the potential for a super app experience. In western markets, those things have evolved completely separately. And so, there's no way to take those independent things and kind of tangle them back together. So, that's why I think this super app strategy is kind of a pipe dream. You wrote a really well-argued piece about that recently. To So, so just to kind of put a stake in the ground, so you're very skeptical of the super app approach that ChatGPT's taking. Yeah, because it creates this chicken-egg scenario now, where you're you have Every retailer already has a very well-designed e-commerce experience. It's built for purpose. Now, you're asking you're asking retailers to start

building a new experience and manage essentially a new e-commerce touchpoint when the consumer behavior hasn't been demonstrated. Why am I going to invest all those resources in until the actual transaction behavior's there? And so, that's the chicken-and-egg scenario. We've actually seen this happen before, and it's failed. We saw Facebook Shops, Instagram Shops, these other e-commerce efforts on social networks, and it never works. Well, we've also seen it, you know, my background being in in a mobile, and specifically mobile gaming, we've seen it with just alternative app stores, right? It just hasn't worked. Not to say that alternative app stores can't work, and maybe there's I think there's an opportunity now with some of the, you know, rules being relaxed as a result of the Epic court cases. But in just in in like the Amazon Fire app store, for instance, right? It just never took off. And I mean, the problem is exactly what you stated. It's, you know, they it was just this dance that you would do as a developer every year, right? So, I was, you know, working with that Rovio as a VP of user acquisition, and so you just Amazon would come to the

headquarters in Finland, and they would try to pitch, you know, Amazon Fire, and you should be on the Amazon app store, and it's it's like, "Well, okay, what are you going to pay us to do that?" Because I can't, you know, take a bunch of engineers and direct them to, you know, spend X months building for your platform. And so then, you know, they'd have to offer some money to make that attractive. And I don't believe ChatGPT's doing that for the app store. I agree, I don't think the I don't think the ChatGPT app store is going to take off. Um there's And that could change over time, but like I try, you know, the one of their launch partners was Zillow. We're sort of, you know, we're being opportunistic with like house hunting right now. So, I'm using Zillow every day. And it was like, well, I actually, if I could go and just use natural language to do house hunting, that would be probably much easier. But it just doesn't work. Like if you go to that app on ChatGPT, it doesn't work. And I don't know if there's a structural issue there, or that's just something that could be smoothed over time. But it did not We I'm not saying like, "Oh, it was harder to use than I thought." It just didn't work. It was utterly dysfunctional. Well, there's a lot of work to be done to get things up to par. So, that's the

hard work in building these new markets. But I think you're hitting a larger point, which is that these are general-purpose apps. They're not built for purpose. And so, I think consumers know to make those tradeoffs. There's a certain moment in time when the general purpose works for you. Might help you narrow some things down, help you in the research process. But once you're deeper into the funnel, at that point, you want the app that's built for purpose. Yeah, and it's not always the case that natural language is just an easier interface, or a more convenient interface than clicking some checkboxes, right? I mean, there's it's not it's sometimes it's longer. It takes longer to type out Like if I'm ordering a pizza, is it longer to like that was the I don't know if you saw this, cuz this to me was just the, you know, the sort of ultimate resignation of this kind of agentic commerce approach to being nothing more than like kind of PR, but like there's this big announcement that Starbucks was integrating into ChatGPT to discover new drinks. I'm like, who's going to use that? But like if I'm ordering a Starbucks drink, is it faster to like click a button or is it faster to type

out I want a grande Pike with a splash of moment? That's not faster. Good apps are already pretty simple. It's a couple of clicks, right? Like the agentic commerce use case always made me laugh because so many of the commodity purchases that people are talking about it being a use case for is like two clicks on Amazon. This isn't a hard thing to do already. So you know, I think the reality is listen, we are seeing search evolve. So we have to acknowledge that but as I said these other apps are fit for purpose. Why do we think that the traffic and behavior is going to move upstream to the LLMs rather than the LLM technology moving on to the site. And so what we are seeing is Amazon Rufus Right.
take off because people know to go to Amazon for a purchase and now Rufus is actually helping provide that assistance along the path to purchase. Yeah, and I do want to be clear here because I mean, you know, we're being a little flippant right now but like in our original conversation, I mean we were addressing these substantive issues. Like the substantive incentive misalignment, the substantive frictions

with agentic commerce. It wasn't a case of like, well, it takes longer to type out than it takes to click. I mean that's that's obviously not always the case. That wasn't That's I don't want to I don't want to I don't want to mislead people who didn't hear the first conversation that like that is the substance of my resistance to this agentic commerce idea. The actual substance, the biggest point that I made in agentic commerce is a mirage is that Rufus will capture the agentic commerce opportunity on site where there is opportunity to like meaningfully reduce friction in commerce through natural language, Amazon will capture that. It won't exist off of Amazon. There's too many places where the incentives break down. And it matters that Amazon's big. E-commerce is not fractured. E-commerce is very concentrated. Advertising is very concentrated. You're not going to see places where independent agents whittle away this engagement from the incumbents because they are probably fundamentally by almost definitionally very concentrated. That was the substance of my sort of

skepticism. It wasn't that, you know, oh start this is a silly use case and Starbucks has an app. That was wasn't it. Yeah, we've seen this history play out, you know, 20 years ago with the dawn of search in general, right? It started with Google and then that search technology moved onto the e-commerce sites. And now we've lived in this steady state for a long time where some of the activity will happen in Google initially and then it moves downstream onto the site. So I think it's a direct parallel. I want to talk about though the actual experience of these, you know, so so just once again kind of want to reiterate here that like our in our original conversation, we were talking about these like very sort of fundamental substantial problems with conceptually with independent agentic commerce. But I do want to talk about what these demonstrated engagements problems were, right? So Walmart noted that it saw three times worse conversion from instant checkout than with clickout purchases. And so that defies I think what a lot of people cite as the advantage of agentic commerce kind of in the abstract. Well, it's just it's you're going to be having more intent as

you discover these products and so there therefore you're going to be more primed to convert and that's actually not what was observed, right? And so what do you think explains that experience? I mean, you know, and I think people might say, well, this is early days. And actually my rebuttal to that is no, it's late. This is late. If what we saw with the instant checkout experience was so disastrous that it opening I walked away from it in a matter of months uh in favor wholesale favor of ads, then it's we're late. It's over. This is it happened, right? But talk to me about that. Why do you think Walmart's had that experience with agentic commerce? Two main reasons jump out to me. The first is you know, we've seen the buy button experiment many times over and consumers have already weighed in with their vote. They don't like buy buttons. They don't like to transact through an intermediary. They tend to like the direct relationship with the retailer. They want to know that they have a throat to choke if something goes wrong. So there's just a lack of trust. There's always some initial lack of trust with, you know, things like financial security and stuff like that. People ultimately will get over those sorts of inhibitions

but I think that's the big thing is that people just don't like to they don't like to transact in those environments as much. Now, the other thing is if if agentic commerce is working as it was proposed, then they're being delivered a single option and I would say consumers actually don't want that single option. It sounds nice and efficient. You actually as a consumer get less satisfaction when you only have one choice. And there's been a lot of academic research, Adam Grant and others, you know, around the paradox of choice. And there's this kind of inverted U curve where, you know, on one end if you have too many choices, it's overwhelming and you end up not liking that result. But if you only have a single option, you also are not satisfied with the purchase because you don't know what you're missing out on. So you can end up with the exact same product but if you have three or four options versus one option, you're going to be more satisfied when you have three or four options. So I think that's a key thing is consumers just don't want that single option at the end of the day. They want to feel confident. They want to build conviction in that purchase

when they make that purchase. Yeah, and I think there were other issues too that they were making the point that this only facilitates one purchase, right? But you know, the goal of an Amazon or a Walmart is to get you to fill a cart with lots of stuff with adjacent related products and it's not just to drive a single purchase. Right. Well, it's right. So again, it's not fit for purpose. So then you say, well, it's not working because you can't basket build and then you start building all these additional bells and whistles and utility and all of a sudden what have you built? An e-commerce site.
Right. So why don't I just click through and go to the e-commerce site in the first place? Which was one of the points that I made. It was that if you want to address the functional shortcomings of this, you're not just building a better discovery engine on top of these retail platforms. You're building a better version of these retail platforms. Now if that was possible, that would have already happened. Exactly. If if you could unseat Amazon, you would have already done that. The you know, it's the prize for doing that is is too great, right? To not sort of incentivize people to try and it hasn't happened. And these companies are too unfocused. I mean, they're all over the place. They're throwing spaghetti at the wall.

What makes us think that they're going to focus enough to actually execute on this? Yes, they have a lot of resources but they've also got to start generating revenue pretty damn quickly or they may not exist a couple years from now. Well, you know, and thus ads, right? But so that that argument it was like people would say, well, you know, look, OpenAI can go back to the trough whenever it wants. And sure it can and it just did. It just completed the single largest private fundraising in history but they don't have as much money as Amazon. Like Amazon doesn't have a lot of resources. Like what are you talking about? You know, these the companies that it purportedly wants to compete against in this use case are far larger and richer and have more AI you know, like machine learning engineers than they do. Plus a massive bank of data which is the actual asset. Yeah. And listen, OpenAI only has a couple thousand employees and you're going up against companies with hundreds of thousands. Well, I I think what I found really interesting about that experience was that it was I think I don't maybe

the most controversial thing I'd ever written, right? I mean, I got more pushback on that than anything I ever wrote related to privacy or just anything else and it was just an exceptional amount of ire that I experienced from that piece and then the you know, the follow-ons which was really interesting to me and you kind of raised this right before we hopped on the call but like I had spoke about this with Ben Thompson because he had had a similar experience with one of the pieces he wrote and his point was that he thinks that it's just the stock market is being propped up right now by AI and there's just like a lot of fear and there's a lot you know, there's a lot of uncertainty and anxiety around what happens if this we just can't prove out a commercial use case here, right? Beyond beyond, you know, what companies obviously integrating, you know, on the enterprise side but like on the on the on the consumer side, what if there is just no robust use case here. Actually I read in Wall Street Journal this weekend, you know, they have the AI adjusted mag seven. So they swap out Tesla with Broadcom. And so if you remove the AI

adjusted mag seven which is Broadcom, Alphabet, Amazon, Apple, Meta, Microsoft and Nvidia, if you remove them from the S&P, the S&P's down on the year. Yeah. I do think that maybe explains a lot of it a lot of the hostility. It's possible. I mean, I so I I took a lot of arrows after my article as well. I had never really experienced that. One of the reasons maybe is that I'm not on Twitter but you know, you see that the reports of what happens with people are anti-Bitcoin or you know, thought that NFTs were BS. And so it it was kind of wild. Um you know, maybe it does ultimately come down to that. Now the whole stock market thing, I mean, first of I don't invest in any individual stocks. I have no horse in this race and I think, you know, you and I as analysts, Ben Thompson as an analyst and actually if you look in in my industry, all the retail analysts almost universally were skeptical of this thing. And I think that the broader market should look at that because analysts if there's one thing that we all do the same, it's we compulsively look at both sides of the issue, right? We're testing our own

assumptions. We're trying to see what argues in favor of this happening, what argues against. And if all the analysts in this ecosystem ultimately come down on the side of being skeptical, I think it's a good time for the rest of the industry to start questioning things as well. It was redolent of the NFT phase, I think. The just this unbridled enthusiasm. Like how could you be against this? This is so obviously the future. And and you know, I kind of like question what kind of future you even were depicting if you think this is where commerce is going to head. So like I've been doing my Prosperous Society series podcast series and in the part two is about agentic commerce and you know, I made the point that look, if you think AI is going to obviate the need to work, why would you want it to obviate the need to use the fruits of this labor to express yourself? If AI is taking my job or doing my job for me, do I also want it to do the thing that I use the benefits of that job for, which is buy stuff? I don't think so. Like and so if

you're saying and again, if you say, well, okay, but only for the staples, only for the gro- and we talked about this last time, only for the groceries and the exit. Well, you can do that now. You don't need AI. You don't need to train a neural network. I mean, the point you don't need to train a neural network to know when I'm going to run out of toothpaste. I brush my teeth twice a day, right? That's not what AI is is bringing value to. And so like you can already do the simple stuff and it's like the big stuff is actually the thing that I take joy in researching and spending time. Like I've been looking at buying a car for like 6 months. Part of it is just cuz like it's fun to do. I like to go through Autotrader and like pick different years and models. It's just it's enjoyable. And that's a part of human expression. If you want to suppress that, I don't think I don't know what we're left with. Right. I we're getting philosophical here, but I completely agree that we there's this whole techno-determinist mindset that thinks that, you know, the technology dictates how we will behave in the future and that essentially we're going to let the bots take over and at some point we just become the humans from Wall-E and we're fat and every- we just sit in front of screens all day. We are

abstracting humanity out of the process. And I would say like I'm pro-humanity. I want to get meaning out of life and I get that from doing the work that it takes to enjoy things and achieve things. Yeah, you know, I'm you know, I don't know that I could be any more optimistic about the impact of AI on society. And I don't know if I could be any more of an AI accelerationist either. But I also think we can pick and choose where we want to apply this and we should apply it to the things that make our lives better and not actually the not apply in ways that strip of strip us of our sense of self. And like that it's weird that think that, you know, people got so upset, just so agitated that I was making the case that like I don't think this is going to be something that people adopt broadly, first of all. And then, you know, second, where it does make sense, it makes sense to tack it on as a feature to Amazon or to Walmart and not as a broad product category. And I guess that one of the reasons people some people got upset is cuz they're building startups in in the space that presupposes it becomes a big new product

category. Yeah. Yeah. And so I mean, I think that's the at the end of the day, those who seem to be most pro-agentic, I found have a huge vested interest in it. As I said, I don't. So I don't care. I'd much rather put thoughtful analysis into it and be directionally correct than, you know, make a fortune on it. Plenty of mobile growth strategies are still built on broken journeys and incomplete data. And it's costing brands [music] big-time. Branch fixes that. It's your AI-powered MMP that unifies your links, measurement, [music] and user experience across every channel to successfully acquire, retain, [music] and engage your app users without wasting spend. Learn more at branch.io. And while you're [music] there, check out Branch's AI search and discovery report covering insights from more than 300 marketing, growth, and digital leaders on how they're responding to the rise of AI search. Okay, let's I want to talk about Rufus. How is Rufus evolving? Amazon moved

sponsored prompts to general availability last month, so that's ads in Rufus. I mean, you know, they're finding new ways to monetize it, right? That I think the last touchpoint we had with any sort of commercial impact on, you know, retail sales was 10 billion in annual incremental sales as a result of Rufus. What's the latest on Rufus? Yeah, well, there's been some good third-party analysis. Sensor Tower has been doing a bunch of it and they did some analysis around the holiday season, which said something like 40% [snorts] of transactions that occurred had, you know, some interaction with Rufus along the way. So one that's interesting, right? It's certainly a key touchpoint, I think, when you have that much activity flowing through that sort of behavior, it's natural that you would find some sort of advertising use case. So we're going to see some sort of evolution of what that right ad format is. But one of the analyses they did that was so interesting to me was they looked at all the different Rufus searchers and they classified them into 10 different consumer segments and how they were using it essentially. So some of those segments were search

assistants, people use a search assistant, research conversationalists, product validators, quick checkers, things of that nature. And if you look at all 10 segments, the interesting thing is none of them were of the set it and forget it variety, where they just type something in and tell Rufus to do the work for them. They're actually doing things to to take them through the middle part of the funnel. Like they're trying to build that confidence to the towards the purchase. So they go in, I think, a lot of times with very high intent. They're more likely to convert. It's, you know, to to what extent is that causation versus correlation? But ultimately, to me, the whole use case of the AI tool is about moving through the middle of the funnel, which is probably the most important part of the funnel and the most undervalued part of the funnel. And so the interesting irony of all this is that the advertising industry right now is built on traditional brand building advertising, where the KPIs are all about things like reach and impressions. And in the bottom of the funnel, where it's clicks and

conversions and ROAS, we've never had a great language around the middle of the funnel, but that's actually where the heavy lifting happens. So that's the interesting dynamic that I'm wrestling with in my head right now is we could be at the dawn of a really new era of advertising, where we get better KPIs and understanding of what moves people through the middle of the funnel. And if that's true, there's a ton of economic value associated with it. But what is that magical metric? What's the that single unifying metric that makes that the whole industry starts to unify around? I think that's yet to be determined. And so that's kind of what I'm watching for right now is how does that all manifest? And from OpenAI's perspective, if they have a mega ad business, I think it's going to be in that middle of the funnel. And so right right now like I mean, let's talk about open ad for a second. They have they can basically do your Yahoo! homepage business because they have a lot of traffic and that's kind of what they started with. They're they can have that search, you know, Google-like click-through business. But the problem

there is if they're actually giving you that one perfect recommendation for a product, then that comes into conflict with whatever they're advertising. So it actually needs to kind of work where people in that are in that early exploration phase. That's where you introduce the ads and that is in the middle of the funnel. But then what does that look like and what are the KPIs associated with that? Yeah, I so I agree. So I had Google's Dan Taylor on the podcast I guess it was last week, maybe 2 weeks ago, I don't remember. But he was talking about that is it's like they've been exploring where to insert the ads, right? And then if you kind of wait for, you know, more substance to have been unearthed in the in that conversation, it's you get better outcomes. And I think that's probably right. I I mean, it clearly is if that's what they're seeing. But that makes intuitive sense to me. I think so my I and I agree that it shouldn't just be a search model. I wrote a piece recently that search is just a a flawed mental model for how these ads should look. I think it should

just be display that's based on behavioral history that's collected. And my sense is this is probably what they're building. But I think the search model, it it there's a lot of reasons, but I think the most obvious reason it it's kind of ill-suited for that engagement profile is because of the questions around, you know, who are you serving, right? Am I serving the advertiser or am I serving the user? And if there's ever any doubt, then you just kind of poison the well. But I mean, that's a very obvious observation. You know, in in terms of like guiding people and massaging that transaction, you know, in the right direction that leaves the consumer really satisfied, I agree. I think that is the opportunity here. But I do want to kind of revisit that 40% metric because I think like you you're underselling it a little bit. It's more than interesting. I think that's astonishing, right? And that's again like why I say it's not too early, it's too late. Amazon won already because if you've got 40% of Amazon's transactions during the holidays flowing through Rufus or being influenced by Rufus, then yeah, they've dominated agent to commerce. I mean, that's you can't

that's insurmountable. That's an insurmountable advantage. Yeah, it's huge. And I think that's the reality. I Now I suspect that it's you work backwards from the transaction. A transaction is going to happen. You are more likely when you are in that very high intent mindset to to get yourself to the point where you can actually feel confident clicking. So I don't mean to suggest necessarily that 40% when you go in is going to be 40% of the transaction. It's probably quite a bit higher during the holiday season cuz we know intent is the highest. That said, if there's no doubt it's going to touch a very high percentage of transactions and be a key part of that path to purchase going forward. Well, it kind of reminds me when, you know, I was getting pitched a lot by AI creative production startups. And I was saying like Facebook's going to own this. Like you can't beat their data, A, but they just have they already have access to this massive, you know, group of advertisers. You're going to have to you're going to have to fight them tooth and nail for that. They have it. It's something that they have access to the customer base. And then

they, you know, flipped the switch, turned on that tool and announced 1 million of their advertising clients using their AI-enabled creative generation tools. And then like I think 6 months later it was 2 million. And it's like, okay, you're a startup competing there. So now you're talking about, well, exclude Facebook as a place to deploy these. And what are you left with? Right? Again, it goes back to that concentration and that that access issue. But then going back to the middle of the funnel thing cuz it it's really interesting. So my sense is like if [snorts] you look at that as the opportunity space, first of all, I agree with you that like we're talking about something new here. Yep. This isn't, okay, you know, SEM that is just has like a different front end. This is something fundamentally new and that's exciting. That's much more more than I think a lot of people are giving this credit for. And also, I don't think you actually exclude the notion here that this is display because it could be display. It could be typical behaviorally targeted display if I'm guiding you through this transaction using world

knowledge and not inserting any sort of influence from advertisers just in the way that I help you figure out what to buy. But, I can use advertising as a way to give you recommendations on adjacents, things that you might fill up the basket with. And that's where I think Rufus is doing. I mean, that's the whole point. It's those follow-on queries. It's like, "Okay, here's I found you the thing that you want to buy and then I think best suits your needs and therefore everyone wins because you complete the transaction on the retail platform. You get the best possible product. You don't have to worry about whether my recommendation was influenced by an advertiser bidding against it or not." And then also, using my knowledge of what people also buy as a result of buying this one thing, I can make some suggestions that are mediated by bids, by ads, as ads. Yeah, I mean, that's where incrementality happens, right? You have related purchases, maybe you upsell people into a more premium version or a larger assortment. All those are places of incrementality, but also, the thing I think about a lot is there's a ton of

advertising value in when you find that person who's in your market, they're in the early exploration process and you introduce your brand, especially if you're a new brand, a challenger brand, just make them know that your brand exists and what your brand is about. And they may not be completing that purchase or intending to move through the path to purchase at that time at all. And I think that's an interaction that gets very undervalued because we don't have that performance data. You don't have the click-through at the end of the day and the attribution. And so, if I'm an advertiser right now, I'm actually looking for all the moments online where that's happening and I don't have to really pay for it. Now, this this creates a huge measurement challenge. How do you start to introduce the attribution or the credit or the KPIs so that I understand that impact is happening and accruing value to my brand even if it doesn't translate to a purchase immediately. I mean, honestly, that's I think something that a qualified marketing team, I mean, that that kind of a brand broadcasting should

always be something that they're doing. And the difficulty in that is measurement, right? But like really good teams, really, you know, sophisticated teams have mechanisms for doing that, but I think it's also like this also sort of is a benefit of these big fully integrated retail platforms because well, they know. They can say, "Your brand was mentioned as a participant in this category." And, you know, maybe not that day and maybe not 3 months later, but 6 months later when that person was in market for that thing, you know, they bought your product. And they have all sorts of ways of testing that to seeing to see if that was actually incremental and if that would have would have happened anyway or, you know, in in assessing the influence that the original sort of just brand ad had. And they are much better positioned to do that than an independent agent who can't actually verify against this long sequence of behaviors. Yeah. And they're going to have to do the research to prove out those effects and then develop the nomenclature around it, develop the metrics and then teach an entire ecosystem. And you know, listen, we this happens when whenever we

have new ad formats, we always have these cottage industries that that pop up around them. So, that's certainly a possibility that happens over time. It's just new and we have to accept the fact that there is probably value there and if we're going to capture that value as an industry, a lot of this new stuff is going to have to take hold. I think about Pinterest a lot in this context because to me Pinterest is almost a perfect advertising channel in some ways because it's, you know, what you're doing on that site is so uh you can draw a direct line from the behavior you're doing to a a product purchase or commercial behavior. And yet, it's been a very undervalued channel for a long time because it actually excels in the middle of the funnel, but it's always gotten under attributed. And so, they've actually had to steer more towards performance advertising just to get credit so then they can capture those ad dollars. So, it's not an easy challenge at all. Yeah, I mean, that's it cuz again, I mean, you're as in I mean, we I don't have any questions for you about Pinterest, but I mean, that's a hobbyhorse of mine because I wrote this piece years ago called The Promise of Pinterest and it's exactly what you

said. I mean, this is a retail discovery platform. Fundamentally, that's what that product is. I'm going there to find things to buy. And it's almost like tragic the degree to which they've just kind of failed to gain traction or failed to grow that ads business because I imagine it is driving a lot of value. The problem is they're subject to the measurement limitations of their clients. And that's why their client base are very large advertisers. It's not the long tail because the long tail can't convince themselves quantitatively of the value. The large clients can. They have the tools. They have the infrastructure. They have the measurement apparatus. The small clients can't, right? And then they can't directly make the case as a meta can or as Google can just through this massive corpus of data that they receive through all of their pixel and the and the CAPI. And that's what I've called small platform syndrome. That's just budding up against that that limitation. But yeah, Pinterest is I just think like there you know, there's some set of

circumstances, there's some alternate reality that we live in where they they've just exploded as an advertising platform cuz they could either prove that or some measurement tool becomes just more widely available. Yeah. And the another perfect example of how our industry, you know, because of data and attribution, we've abstracted away the humanity of it. And so, when I like sometimes I I just have to remind people like, when do we stop being marketers? Like you can't tell me that seeing an ad in that context is intuitively valuable to you. Figure out what the metrics are that help articulate that if that's what you need to drive investment, but we need to start thinking a little bit bigger picture because the metrics actually are increasingly steering advertisers in the exact wrong direction. Those metrics are increasingly gamed and they're steering us away from I think a lot of the most valuable marketing channels today. Let's get [music] real. Half the time, you're not sure if your campaigns work. You just kind of hope they do. Attribution numbers look good, but what if those [music] conversions would have

happened anyway? Incremental helps you find out for real. No tracking people, no guessing games. Just clear answers on which campaigns actually move the needle. If you want to stop guessing [music] and start knowing, go to Incremental. Because good marketing isn't about getting clicks, it's about getting results. Check them out at incremental.com. That's i [music] n c r m n t a l dot com or click the link in the show description. You shared some slides with me that I guess it's like a keynote that you give. In in one of the slides that you shared, you had indicated that AI search tops out at roughly 2% of referrals for large retailers. Do you think the steady state of call it agentic commerce, agentic discovery, AI-assisted commerce, do you think the steady state is a discovery tool for clickout? And if so, how can these platforms capture value outside of their own ads and through subscriptions, right? So, I'm now I'm talking about like the independent side. If it is just clickout purchases, if that's the steady state, how do they extract value from that?

Yeah. Well, I think it is that is the steady state. You know, I expect those numbers will go up, but it's a new channel, right? We've added channels in digital over the years, so this becomes a new one. One point is just to show right now, it's not a huge referral. Not a huge part of referral to top retailers. Part of this was kind of speaking to the existential threat that GenAI was supposed to be and that it was going to pull all this traffic upstream. One, it's not. It goes anywhere from 0.1% to about 2% for top retailers. But 60% of top retailers' traffic is direct anyway, so that's not going away to begin with. And then what's interesting is where you see the percentage is higher, it's like Best Buy, Home Depot, Lowe's. These are higher very high consideration categories. So, again, when you have really high involvement, you're much more likely to do some of that exploratory research and then end up clicking through. Now, if it was true that higher GenAI activity was eating into e-commerce traffic, you would see

those higher referral retailers declining in traffic, but you're not. You're seeing the exact opposite. You actually see a positive correlation between the share of GenAI referral and traffic growth to retailers. So, but there's probably a better argument that GenAI and having new forms of search is an incremental behavior and something that actually will add to e-commerce growth if anything. Another slide that I found to be really fascinating was just the search share by searcher category, right? So, you had the early adopters at like 41.4% as the most recent data point in the series, but that was flat. It was flat for three three months running. And then you had the all searchers was also flat, but it's it's it seemed to be kind of it was on a it was on a sort of upward trend, but it was flat for the last two, essentially flat for the last three. I mean, is this the step I mean, cuz I would imagine, first of all, I would have expected that the early adopters' number would have been higher if this truly was like this new sort of this new consumer

behavior that was going to be established as a norm. That should be more than half, I would assume. And then but then for this to have flatlined already, again, this is kind of reinforces my point. Like this is not too early. This is too late. It's over.
Yeah. Like Amazon owns it. Walmart owns it with Spark. I mean, like all these retail platforms getting 2% of their clicks are from Apple. Who cares? Like they'll build their own agents. They're going to integrate directly. They're going to just get you to the transaction and again, service that messy middle. But talk to me about that. Talk to me about that the flattening. Yeah, so this is looking at LLM search. So, so basically the share of search and it's this is from like early 2024 till a kind of late 2025 is this data and so what it's showing is it goes from about 2% to 6% overall of searches which is non-trivial in a a massive market like search, but it's not overtaking the market. There's a subset that we look at here of early adopters and basically this are people who are already searching on LLMs back in early

2024. And that you see a a sharp uptick from about 8% of searches to um it goes up pretty quickly into the 30 and 40% range. But then as you imagine it starts to level out at around 40% in late 2025. Actually, there's another a follow-up slide to this one that does a cohort analysis to say, well, is this just an early adopter phenomenon or are newer cohorts doing the same thing? And the newer cohorts are sort of doing the same thing. It's not as sharp of an uptick to like 40%, but it's still fairly pronounced. So that suggests that you know, activity should be moving in this direction for search overall. Now the interesting thing is though it did flatline and that flatlining has continued. What I think a lot of folks do not understand is that if you're to look at ChatGPT data going back to around July or August of last year, so like 8 months time at this point, it has not budged in terms of weekly active users. It has not budged in terms of total sessions.

Like it's flat in the US at least and you know, maybe incrementally growing globally. But this is not a hockey stick phenomenon anymore. It started out that way, but then it hits a natural limit. And again, I've seen this growth curve before where everyone assumes that the exponential curve goes up into the right forever. And it was the early days of Twitter. Twitter had this huge hockey stick and then all of a sudden it came to a screeching halt and it just went flat for a while. Then it started to build in a a very steady incline over a longer period of time. So I expect something similar will happen here with ChatGPT, but we just need to acknowledge the fact that we are no longer in this exponential growth curve. Well, I have a lot to say about ChatGPT self-reported wow number. So first of all, wow is a non-standard metric. I mean, people say, oh well, but it makes more sense to I mean, I've heard this criticism and my criticism is that wow is a more natural reflection of the usage, but that's not the point of a metric. The a point of a metric is to be

comparable. It's not to better capture whatever you think the of an externally reported metric. If they're using that internally because they think that better captures behavior patterns, fine. But that's not what that's not how they're using it. They're using it as an external reporting metric and it's non-standard. And then, what's the point of adopting a non-standard metric is to obfuscate the actual underlying dynamics, which is I think why they've adopted it, right? Now the other thing here is that they said 900 million wow when they announced the fundraise in February. You think they would have updated that when it hit a billion. They would have. That would have been a headline. They haven't. Obviously, the growth is slowing down. Because it was I want to say it was 850 was the most it was the release before that and that was just a few months prior. Right? I mean, they'd been engaging in a steady drumbeat of metrics reveals and now they're not. And a billion is a pretty big milestone, right? Now maybe they're above that now and they just were waiting for something to pair it with, but the big question is what's the

wow mau ratio? Third-party data speaks to this. So it it's pretty similar trends whether you're looking at mau, wow, or dau. They're literally, you know, in parallel. Some of them might show, you know, very minor increases. You're if depending if you're looking at weekly, there's even some suggestion that there's you're seeing some declines and some of those declines, you know, very incremental, but you know, seem to align with post-Super Bowl and some of the Right. you know, Anthropic catching gaining steam. So but yeah, I think no matter how you cut it, right? It's not a it's not a strong growth story anymore. And I think we're seeing that both from the external reporting and and what third-party data is showing. Well, that's why you would use wow. You would use wow because it's more sensitive to big swings. That's why you would use it. And even if it's backward-looking wow, if you're averaging out these very large numbers because there was a model release, then that's why you would use it because you could time it to show higher number. You could time it to show number goes up. Now there's nothing inherently wrong wow as a metric, especially an internal reporting metric, that's fine. If that is more aligned with, you know, usage

patterns, you should use that. But in my point my and I wrote a piece about this a while back, but like my point was you can't compare it with anything because no other company uses wow as an external reporting metric. And also, it's going to be more susceptible to swings. And so since they're not doing it on a regular cadence, they're doing it after model releases or they're doing it after some other big event, it seems like they might be using it to game the releases, right? And that was my issue with it. Just lacking the comparability, I think is a key point. Although, I could also argue it maybe it makes sense because long-term this is behavior that should very much be a daily habit for people. So long-term you have to be reporting in terms of daily active users. So in the interim as they're still building that user base, maybe it's fair to say weekly is is a a decent metric for now. But yeah, I think the the bigger factor is probably just lack of comparability. Okay, so I want to get to the I want to get to performance TV. So so you wrote a great piece recently, will CMOs finally wake up to retail media in 2026? And you sort of make the case that performance TV and in-store retail media are the two

significant new opportunities in 2026. Walk me through both of these cuz I think I misinterpreted what you meant by performance TV at first, but just kind of explain what both of those are. Yeah, so performance TV is a lot of times people think it's shoppable TV. So whether I have a QR code or an add to cart or I have to click on my remote to buy something right away. I think that's always going to be a very tiny behavior. When I say performance TV, I mean basically you have TV ads, you know, CTV ads specifically, underpinned by closed-loop targeting and attribution. So think about primarily what what Amazon is building here with Prime Video and also now that Amazon has plugged into all the other ad-supported CTV networks, that's performance TV. And that's a big market opportunity and will really change the dynamics of TV advertising. Cuz you still have the benefits of TV ads which have always been about brand building and reaching the right audiences, but now you can actually see if they're driving sales as well. And so there's all sorts of interesting dynamics that come into play when you force fit performance metrics

into this equation. In-store retail media is just we're seeing a rise now of in-store digital ads into physical store environments. CVS has moved in this direction. A lot of regional grocers and now Kroger and Albertsons. This is a big thing by the way in every other market but the US, we're very late to the game here. But it has a ton of potential. And the reason why these are I think increasingly in the crosshairs of the CMO is a lot of CMOs have kind of pushed retail media to the side as just this e-commerce performance channel and it's not what they've traditionally cared about. They're brand builders, creatives, brand marketers at the end of the day. And my argument is the CMO of the future and advertising of the future, you've got to be able to do both. You have to understand brand building and performance at the same time. Have to be able to walk and chew gum. And these are the almost the two perfect venues for the CMO to care about because they're scaled media that can do unbelievable brand building with great

creative. [snorts] And I think in-store digital is a creative opportunity unlike we've ever seen because you can,
[clears throat]
you know, have executions happening kind of in three dimensions across the store in really unique and interesting ways. And also they're going to drive performance. So how do you blend these disciplines and be able to think about it and have reporting that really captures the big picture of reaching audiences at scale, how valuable are these audiences, who are these audiences? And then within the short-term view of the world, you know, are they is it leading to incremental sales as well? It can't be one or the other or I think you you miss a lot of the value. Just for my own personal edification, but I imagine not everyone's as familiar with this as you. What formats do these take? I mean, is this is this genuinely a digital billboard in a grocery store or what what does it look like? Yeah, I mean, think of TV screens. You can have them at the checkout aisle. You sometimes have these, you know, 6-ft tall or so digital screens maybe at the entrance. You'll see TV screens at the

pharmacy counter or the deli counter. Um you can have some around the periphery of the store. I think the end cap at the end of the aisle is probably the most prominent place, maybe the highest impact format in the store. So it's just increased digitization and you know, you have to execute in a way that does not detract from the CX. So I'm also very concerned about introducing what I call the physical store equivalent of pop-up ads. You don't want to have ads shouting at you. It doesn't want to be promotion. But if you have high-quality creative dotting the store in interesting ways, it can actually elevate the store experience, modernize, and you know, introduce some content to the store experience that that enhances things and will also introduce people to brands, new products, and drive sales. And and this is bought through the just existing RMN infra. So like you Target's around L and Albertsons is through Albertsons and Kroger's is through Kroger. You're just buying that through their existing interface. Yeah, that's primarily how it's going to happen in the future, but there's also program There's dollars driving through

programmatic right now as well. So think of it as like this is an extension of programmatic digital out of home. It's just in a unique context. Is it like I'm thinking like ad quick or is that is it like trade desk? Yeah, I mean, it could be through both. Who whoever is is driving those dollars through programmatic. So that's where most of them come today. I think it probably makes more sense. I'm kind of arguing that we should go back to the future with in-store ads and plan and execute them like we used to do for linear TV buys cuz you can kind of plan by day part, by geography, and all these different factors that I think are much more of a scaled ad buy and yet there's such an inclination to just buy everything programmatically. I just think you want to take a lot of care of what the creative is and how you execute it in store. You can't just throw any ad up there and expect that it's going to work. You have to make it native to the environment. This is how every new medium emerges where you have, you know, you just import ads from the last thing into the new thing and it doesn't really work that well at first and then as the

ecosystem builds around it, you get better creative and you start haven't realized its potential. I'm going to open a big can of worms here. I know we've got like 3 minutes left. So so feel free to give me the abbreviated take. But where is Shopify sitting in all of this? So we've got, you know, you were talking about performance CTV. We're talking about agentic commerce. I mean, what is Shopify's role here? Because I my sense is like, you know, you've got this massive retail platform. They've been kind of slow to really embrace advertising. I mean, they haven't had product and they actually just launched their sort of, you know, their network essentially. I had the the director from Shopify on. She didn't like the fact that I called it an ad network, but it I mean, it is. But like, where do they kind of sit right now in the ecosystem? What's your take on their positioning? Yeah, I mean, it is an audience network at the end of the day. I mean, they listen, they have a huge swath of potential advertisers with all these D2C brands. So in that sense, they have a lot of potential and could be a front door to a lot of advertising. It's one It's a platform I'm scratching my head at a lot of the times. I'm not sure that they've fully executed to their

potential. With respect to CTV, I mean, listen, D2C brands are looking to diversify beyond Facebook and Google today. No doubt about it. So CTV is probably going to be the next frontier in some respects. But it's not going to be easy necessarily. They have to get access to good inventory and inventory that's going to perform. I worry a lot about CTV ad fraud. So if you are buying ads against fraudulent inventory, then it's not going to perform from an incrementality perspective and the investment won't be there. And the other thing is CPMs are high for CTV. And so inherently, that kills ROAS. Like ROAS is not a good metric, but CPM high, ROAS low. And so the typical direct-to-consumer advertiser will just optimize out of that. And so or what they end up doing is they buy retargeted ads for TV. And that's the only way to make ROAS pencil out. So I worry that's what's going to happen. It's not really incremental and it may not drive growth for their business over time. But can Shopify play a role there?

Absolutely. Andrew, I appreciate you coming back on the podcast. I was pleased to be in the trenches with you, you know, in in the frenzied 6-month-ago long long bygone era of the agentic commerce hype cycle. But appreciated you you've been at my side. How can people find you on the internet? How can they consume your content? Where do you exist on the internet? Yeah, I'm pretty active on LinkedIn, not on Twitter. And you can check out my Substack. It's mediaadsandcommerce spelled out.substack.com. Thank you, Andrew. Thanks, Eric.