Latest / Retail Media Breakfast Club / Is ChatGPT's Instant Checkout 'Really' Agentic? And Does It Actually Matter?
Transcript
- Is ChatGPT's Instant Checkout "Really" Agentic, And Does That Actually Matter? === [00:00:00] Kiri Masters: The retail industry just loves a good semantic debate. Right now, the hot topic is whether open AI's new instant checkout feature really counts as [00:00:10] ag [00:00:10] Kiri Masters: agentic commerce, or if it's just a glorified app [00:00:15] integration. The argument from Definition Purists is that Chat t's instant checkout integration with Walmart and Etsy doesn't really qualify as agentic commerce because it's reactive and guided by human prompts and [00:00:30] confirmations at multiple steps. [00:00:32] Meanwhile, OpenAI is already calling their system. The agent commerce protocol and retailers are scrambling to integrate, and consumers are starting to normalize AI assisted shopping [00:00:45] at scale according to some brand new research from Klaviyo. Which was released just a few days ago. 78% of consumers have already used AI for shopping or product research in the past three months. [00:00:56] But here's my take. While we're busy arguing [00:01:00] about definitions, the ground is already shifting under retail medias. Profitable foundation, and that is what we should really be talking about. [00:01:09] [00:01:10] Kiri Masters: First, I'll just take a second to explain these definition [00:01:15] wars. The industry has split into two camps. There's the purists that insist that true agentic commerce requires proactive autonomous agents. Here is the definition of agentic commerce from MasterCard. [00:01:27] They say [00:01:28] a [00:01:28] Kiri Masters: agentic commerce is a new form of online and [00:01:30] mobile shopping in which an AI agent closes the loop. Or completes tasks for a user, such as searching for items, comparing options, and making a purchase with limited or no manual inputs needed from that user. [00:01:43] So that's one of several [00:01:45] from bodies like Salesforce and BCG that. Paint that view that there has to be really limited involvement from a consumer. My friend and retail industry analyst from E-marketer, Sarah Marzano, says that [00:01:57] in her book, instant Checkout doesn't meet [00:02:00] the proactive or autonomous definition enough to qualify. Here's what she told me. What we're seeing with the checkout integrations from Chat GPT, and Walmart, and Etsy, et cetera, doesn't check any of the boxes for AI capabilities as the experience is end-to-end [00:02:15] reliant on the human user's prompts and guidance. [00:02:18] Now, the other camp may not necessarily disagree with any of that definition, but I call these the pragmatists, and they're arguing that any AI assistance in shopping [00:02:30] represents progress toward. True [00:02:32] ag [00:02:32] Kiri Masters: agentic commerce and the net. Not getting hung up on the definition, but seeing it more as a continuum of autonomy and. [00:02:42] What I wanna call out, which is difficult to explain on a [00:02:45] podcast, is the agentic shopping levels of autonomy framework that Scott Wino created on his excellent newsletter, retail Gentech. Which defines five different stages of autonomy. By his definition, we are at level three where [00:03:00] AI can help you find products and even complete transactions. [00:03:03] But it is not making autonomous decisions. [00:03:05] But look, here is what is getting lost. While we're arguing about whether or not instant checkout is truly agentic, even this not really [00:03:13] Ag [00:03:13] Kiri Masters: Agentic instant [00:03:15] checkout is capturing transaction data that used to belong exclusively to retailers. So when open AI processes a Walmart checkout, both companies now have that purchase signal. [00:03:29] This means that [00:03:30] Walmart's moat of consumer purchases. Which is a huge moat when it comes to their retail media business. [00:03:40] That foundation just got a whole lot shakier. [00:03:45] Shoppers don't think in silos. They bounce between online and in-store without skipping a beat. Acosta Group's. Connected Commerce agency [00:04:00] helps brands unify trade and retail media budgets to invest where shoppers are most likely to convert, whether that's on mobile or in the store aisle. Learn more about acoa Group Connected Commerce at ACOA [00:04:15] Group and let them know Retail Media Breakfast Club sent you. [00:04:18] That's A-C-O-S-T group. [00:04:24] Kiri Masters: Let me connect this back to an earlier essay that I wrote about retail media's economics. So I [00:04:30] outlined three main revenue pillars for retail media networks, number one is onsite advertising, and this is the lion's share of retail media ad spend as we know it. An estimated 80% of retail media ad spend [00:04:45] is onsite advertising, so sponsored product ads, banner ads on a retailer's owned and operated properties, so that business typically carries [00:04:55] it [00:04:56] Kiri Masters: 70 to 80%. [00:04:58] Margin. So [00:05:00] it's very lucrative for retailers and it is the majority of retail media overall ad spend. Other pillars of retail media include offsite. This is the audience extension capabilities that a retailer has. [00:05:15] They. See all of these signals from consumers, both actual purchases and research patterns and browsing patterns, and then they syndicate that audience data to. [00:05:26] Advertisers so the advertisers can activate against those audiences, and they [00:05:30] might do that with a partner like the Trade Desk, for example. Now, because of all of the intermediaries involved with this, it's less profitable for that retailer, let's say around 40% profit margins. And it makes up a much smaller percentage of [00:05:45] overall retail media ad spend. [00:05:46] Somewhere between 10 to 25% of retail media ad spend is this offsite, uh, audience extension. It's going vary dramatically by retailers. Some [00:06:00] retailers don't even do any offsite, advertising, but this is overall [00:06:03] a. [00:06:03] Kiri Masters: across the entire industry. And then the third pocket is something that gets, you know, it's getting a lot of attention these days, but it is still super, super small. [00:06:12] And that is the in-store, [00:06:15] digital, programmatically enabled in store. Retail media. So digital displays, uh, digital end caps, things that can be bought programmatically rather than your traditional trade marketing, you know, pop-up displays and things like that. So [00:06:30] that accounts for only one to 3% of overall retail media ad spend. [00:06:35] So with all of that in mind, understanding for the. Many of these retailers, their media business is a significant driver of their overall, uh, [00:06:45] gross profit. And a number of retailers have sort of come out to say that retail media is a very substantial portion of their total profit at the end of the year. [00:06:54] So my point in the original essay was that. [00:06:57] Ag [00:06:58] Kiri Masters: Agentic Commerce puts that [00:07:00] business at risk. [00:07:00] And I called out that the onsite media was most in danger Now. [00:07:06] Long term, that is still actually true. But as we start to look at the instant checkout, sort of intermediate step towards [00:07:15] truly agentic shopping, what's become clear to me is that the offsite retail media, um, revenue stream is at more immediate risk from where we're at right now. [00:07:26] It is still at high risk in the long term, but this actually might [00:07:30] be the first domino to fall, if you will. I've put a chart up on the blog that accompanies this podcast episodes. Probably easier to check that out if you want to get into more of the details of the three different stages of [00:07:41] ag, [00:07:41] Kiri Masters: agentic commerce, and then you know what is actually gonna happen to each of those revenue [00:07:45] streams. [00:07:45] But the long and short of it is. [00:07:46] As consumers start using agentic commerce, whether it is AI assisted checkout or true agentic, that retail media is at risk here. The only form of retail media that is not [00:08:00] susceptible to this is the in-store retail media and traditional trade marketing ad spend. [00:08:06] ~Because~ [00:08:06] ~yes, I still think ~because yes, I still believe people are going to go to stores and engage in in-store shopping behavior as well. That [00:08:15] is something that Agen shopping really isn't going to replace. [00:08:19] It's primarily the e-commerce portion that is at risk here. [00:08:23] So there's a lot here. I want to call back one really interesting story back from 1997 when Amazon. [00:08:30] Made its IPO. Barnes and Noble sued Amazon over its claim that it was the world's largest bookstore. And Barnes and Noble's argument was that Amazon wasn't really a bookstore, it was a book broker, so they couldn't call [00:08:45] themselves a bookstore. [00:08:46] And isn't this crazy? It was a lawsuit, uh, over semantics trying to take down Amazon when. That was clearly not Jeff Bezos's end game. He was moving beyond books and [00:09:00] building the everything store. [00:09:01] so look, [00:09:02] Kiri Masters: I think that sometimes with these debates over definitions, it means that we're having the wrong conversation instead, particularly retailers, if you're listening. [00:09:12] Really need to start thinking about what [00:09:15] happens when some of these transactions start flowing through AI intermediaries. Who also own that purchase data. You might still be the merchant of record during this instant checkout era, but another party is going to have that [00:09:30] same transaction record and ostensibly be able to sell that audience extension capability to the same brands that you are. [00:09:39] This is the right time to start thinking about what impact a gentech commerce is going to have to retail [00:09:45] media businesses in the long term. And start putting together a game theory analysis of where this is going to place you in the long term. Wrapping up here. Today's instant checkout might not meet the technical definition of benent commerce, but so what? [00:09:59] [00:10:00] It's already threatening data, exclusivity and conditioning, new shopping behaviors. By the time we get to level five retail media networks, still debating definitions are gonna find their high margin businesses already disrupted. [00:10:12]