Latest / Retail Media Breakfast Club / Agentic Shopping Looks Crazy. So Did Self-Service Supermarkets.
Transcript
- Agentic Shopping Looks Crazy. So Did Self-Service Supermarkets. (Thurs Oct 9) === [00:00:00] Kiri Masters: Memphis, Tennessee, 1916, Clarence Saunders is about to do something that everyone thinks is completely insane. For decades, grocery shopping worked the same way you handed your shopping list to a [00:00:15] clerk who fetched your items from behind the counter. But having someone wait on you wasn't as convenient as it sounds. [00:00:22] Clerks, short weighed products, slipped in, spoiled produce and customers had to wait around for chatty Cathys ahead of [00:00:30] them in the queue. The clerk would scribble your total on the back of a paper sack and add it to a credit account that was floated by the store. The system was inefficient and often corrupt, but it was all anyone really knew. [00:00:44] Clarence [00:00:45] Saunders asked a question that seemed heretical. Why do shoppers with lists need clerks at all? His answer was Piggly Wiggly, the first self-service supermarket. He patented a one-way aisle [00:01:00] system that forced customers to snake through the store. He charged 3 cents for baskets, no clerks, no deliveries, no phone orders. [00:01:10] Critics mocked, not just the silly name, but the absurd idea of [00:01:15] customers waiting on themselves. To make it all work. Saunders introduced tools that seem obvious now, but were revolutionary then and indeed he patented these concepts, standardized price tags that eliminated haggling, adding [00:01:30] machine tape, that printed receipts and customer facing scales. [00:01:35] These innovations created transparency and trust in an industry where that was lacking. Ply Wiggly exploded from just one store in Memphis in [00:01:45] 1916 to 393 locations just three years later. The biggest grocer at the time a and p had 4,000 stores, and they were forced to abandon the clerk model and copy Saunders' self-service [00:02:00] approach because soon consumers were demanding this model. [00:02:05] Now, the lesson here isn't that Piggly Wiggly conquered the world. Managerial missteps later sank the company. But the lesson is really that the [00:02:15] future usually doesn't arrive by consumer request. Shoppers in 1916 didn't demand self-service. They didn't even know that they wanted it. It came from someone re-imagining inefficiencies that everyone else had learned to [00:02:30] tolerate. [00:02:30] [00:02:32] Kiri Masters: Now with that little history about self-service supermarkets looking crazy at the time. I'm gonna use that as a backdrop to talk about agentic shopping [00:02:41] And why it's important to not just cherry pick failures in retail history, but [00:02:45] also look at successes. And this was inspired by an article from Andrew Lipman, who is a retail analyst that is widely respected and, um, I'm friends with Hello Andrew. [00:02:56] And we have debated this topic. More than once in [00:03:00] person. Um, but he did make a thoughtful case on his blog for caution around ag agentic commerce, and he made a number of points. But the one that I'm going to address today is around. [00:03:13] Examples where retail [00:03:15] technology was overhyped, and he calls out a number of examples, including voice commerce, which never took off. Metaverse commerce, which never took off. Live streaming commerce, which never took off outside of China and shoppable tv. Um, oh and, [00:03:30] And IO OT or Internet of Things Commerce. [00:03:32] So each of these promised to revolutionize how people shop. Each generated breathless hype and each ultimately proved inconsequential to the broader retail landscape. And this point is [00:03:45] valid. The retail industry, like any industry, is littered with a graveyard of shiny objects, transformative technology that never really took off or transformed anything. [00:03:56] And given that track record skepticism about new [00:04:00] things like AI shopping agents is reasonable, but here's where I think the skeptical view risks missing something important. It treats all failed innovations as equal when history suggests some supposed failures were [00:04:15] actually just too early. [00:04:17] So let's jump in with a few examples here. Webvan tried to reinvent grocery delivery in 2000 with automated warehouses and its own fleet. It scaled too fast, [00:04:30] burned cash and collapsed. Cementing the belief that people don't want grocery delivery. [00:04:35] Except web van wasn't wrong about the destination. It was just wrong about timing and execution. The model was capital heavy. [00:04:45] It was ahead of consumer behavior. It was pre smartphone, and they built expensive fulfillment centers where there was insufficient population density. [00:04:55] Now, 20 years later, Instacart piggybacked on existing [00:05:00] store infrastructure and used gig shoppers instead of building warehouses. And by now, delivery services like Instacart, door Dash, Uber Eats are ubiquitous consumption behavior. [00:05:12] So it wasn't that people didn't want [00:05:15] this, it was just done in the wrong way. Another example, you're all familiar with this one, amazon.com. Launched in 1995 as a bare bones online bookstore ~and,~ [00:05:25] ~and early detract. ~Early detractors argued that it wasn't a bookstore at all. It was a book broker. In fact, there was a [00:05:30] lawsuit from Barns and Noble right before Amazon's IPO disputing its claim of being the world's biggest bookstore. But the Widert sentiment, especially in the early days, was why would people want to buy books without browsing? [00:05:43] Shelves. It was such [00:05:45] a weird concept, and just e-commerce in general then was extremely nascent. A lot of concern about putting your credit card on the internet. How can we trust these websites? Who are these people? You know, some of these arguments that [00:06:00] are ringing in our ears right now. [00:06:02] But we all know what happens next with Amazon. Massive selection, expansion, across every category, the birth of ratings and reviews, and the consumer behavior that built [00:06:15] around checking ratings and reviews, prime logistics. So what happened is a niche use case, which is buying books online. Became the default way to buy nearly anything, and all these improvements that were made over time ultimately [00:06:30] remove friction for shoppers and built entirely new habits and expectations. ~Pattern early, sorry.~ [00:06:35] ~So here's the pattern. Early versions of this tech. So here's the pattern. Here's the pattern to tie everything together. So here's the real ~Early versions of these technologies can expose real consumer needs, but sometimes the execution is wrong or the missing technology prevents mainstream [00:06:45] adoption. Shoppers don't think in silos. They bounce between online and in-store without skipping a beat. Acosta Group's. Connected Commerce [00:07:00] agency 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 [00:07:15] ACOA Group and let them know Retail Media Breakfast Club sent you. [00:07:19] That's A-C-O-S-T group. [00:07:24] Kiri Masters: Now interestingly about TikTok, you may recall that TikTok was formally [00:07:30] a music app for teens called Musically. It was like a lip-syncing app, and it was just something for teens. Super limited. What's the big deal here? And then it became. TikTok, the algorithm was enhanced. It [00:07:45] became much more , than a lip syncing app and short form video became the global content format. [00:07:52] The use cases for TikTok exploded. Once the product actually matured, what looked very [00:08:00] narrow and gimmicky at the time. Became a cultural default. And that's where I can tie this back to today's agentic shopping experiences. You may have seen, I did a video showing how, chat GPTs instant checkout works and [00:08:14] [00:08:15] i'm putting it out there. I was the first person in the world to buy a lavender scented candle on chat, GPT. Um, if anyone wants to, uh, dispute me, uh, let's compare notes. But listen, I'm the first to admit that experience. [00:08:27] It was a little clunky. It was a [00:08:30] little ugly. It wasn't a, a joyful and like really cool experience. So everyone can agree that the technology here is nascent. Um, there's very few things that you can buy through AgTech shopping. There's very [00:08:45] limited use cases for this. We're still very early on, so could agen shopping end up like voice commerce, finding only very niche use cases. Yeah, maybe, and maybe the skeptics end up being right on this, but here's my concern about dismissing [00:09:00] it out of hand. We can't make predictions about consumer behavior based on technology that is still evolving rapidly, and I have a few. [00:09:10] Consumer surveys, which admittedly should be taken as directional at [00:09:15] best find that consumers are generally interested in the idea of a agentic shopping partic, particularly in more mundane categories like grocery shopping, but also in fun categories like apparel. So. Here's the big question. What should we be [00:09:30] looking for this time around to know for sure if agen commerce is different from the graveyard of failed retail technology? [00:09:38] Well, I I have a few ideas on what needs to align, and we're already there with some of them, but the jury is still, frankly out on a few [00:09:45] more. So number one is the technology itself. We need to have a few things in place here. It needs to be a more immersive experience than just the clunky text interfaces that we have right now. [00:09:56] We need to have better persistent memory so that [00:10:00] our AI agents really understand us from session to session. They understand our preferences, our goals, our routines. , We need. Agent to agent protocols that allow our agent to talk to a retailer's product catalog and [00:10:15] inventory availability more reliably. [00:10:17] There's still a long way to go on the technology front. Second is, is there enough pain in the existing solutions? And this is something that I wrote about on Tuesday where I [00:10:30] really do think our current. Mode of e-commerce is kind of a crappy experience in many ways, and we've kind of just gotten used to it, just like those shoppers at Piggly Wiggly who just got used to waiting in line and get used to all of the waste in [00:10:45] that ecosystem. [00:10:46] So, , go back and check out what I had to say on Tuesday. I think that we're spending way too much time, price checking, comparison, shopping. Between retailers and these are all things that agent commerce can solve for us. So I do think that we're [00:11:00] at least part of the way there with there really being pain in the existing solution. [00:11:05] Number three is distribution. It needs to be frictionless. We don't want more devices. So, and this is the big difference with, uh, voice commerce is, [00:11:15] um, you needed an extra device, but with. Ag agentic shopping. It's already on these phones that everyone is addict addicted to, so I think that that is particularly helpful. [00:11:25] Number four is the economics. It needs to be an attractive [00:11:30] commercial proposition for several parties for the retailer. For the consumers, for the platforms, and one day maybe advertisers, everyone has to feel like they are winning and getting something out of this. And so I think [00:11:45] that we're starting to see this materialize with Chat GPTs App Store, which was just recently announced and we've already seen Instacart, Uber, DoorDash, target. [00:11:55] Already announcing that they are building apps for the chat GBT [00:12:00] Marketplace ecosystem. And number five, which is still missing, is cultural readiness. Trust in algorithmic delegation isn't universal yet by a long shot, especially among older shoppers, there is still a, a long ways to [00:12:15] go in terms of can I trust an agent to buy for me? [00:12:17] Can I trust putting my credit card details into an LLM, , what that payment ecosystem looks like. We are so early days with this. It's going to take a little while, but this is one variable as we've seen from these examples that I shared [00:12:30] before, once it hits, it shifts really, really quickly. [00:12:33] All right. Let's go back to our story about Piggly Wiggly. Clarence Saunders, the founder of Piggly Wiggly, didn't wait for shoppers to request self-service. [00:12:42] He saw inefficiencies that everyone else [00:12:45] tolerated, and he reimagined the entire system. The result looked irrational and weird until it became inevitable. I am not arguing right here that agen commerce is going to transform every aspect of retail and the [00:13:00] shopping experience. I'm arguing that dismissing the possibility based on today's clunky implementations ignores how often these irrational innovations. [00:13:10] Found their moment. Once the surrounding technology matured, the [00:13:15] skeptics might be right. History is littered with technologies that never escape the hype cycle. But history also shows us that some innovations start in a single store in Memphis looking absurd to everyone who hasn't seen what's possible yet. [00:13:28] Here's the thing, we're all [00:13:30] speculating about a future None of us can predict with certainty. But given how often the obviously irrational became obviously inevitable, I'd rather prepare for the possibility than assume that the current limitations reveal permanent [00:13:45] constraints. [00:13:45]