Week 14 - ChatGPT is now a storefront: Shopify's AI shopping play, Amazon's delivery bets, and why post-purchase is the new battleground

Max: Welcome to the Checkout.Points, your quick dive into this week's e-commerce buzz with me, Alex, powered by Blicket. These 20 minutes pack the latest trends, news and insights. Let's navigate the digital marketplace together. Ready? Let's go! Rachel: All right, welcome back to the Checkout Point Week 14. Let's go. Speaker 3: Happy to be here, Max. This week's lineup is genuinely stacked. Rachel: Okay, so you know how ChatGPT has been slowly creeping into everything? Speaker 3: Mhm. Rachel: It's now a shopping channel, like full-on buy stuff without leaving the chat shopping. This is fundamentally different from anything I've tracked, and Shopify is at the center of Speaker 3: Wait, Rachel: it. Speaker 3: so you're browsing, you're chatting, and you're just... Purchasing? Rachel: Yep, and Rachel, here's what actually matters. Your product data is now the marketing asset. Feed quality, attributes, copy clarity—that's what the algorithm's actually learning from. Speaker 3: Okay, that's a big operational shift. Speaking of measurement headaches, AI ad channels are growing fast, but giving marketers almost no performance data. You're basically doing blind spend, which I've seen tank margins before. Rachel: Oh, that's a nightmare. Speaker 3: Right? And on top of that, Visa's research shows consumers still want recognizable trust signals at checkout even when an AI is doing the buying. I've seen this play out in CAC data. Brands that nailed the trust moment had way better repeat rates. Rachel: So you're optimizing for a robot and a human at the same time. Speaker 3: Basically, yeah. Rachel: Cool, cool, no pressure. Speaker 3: We're also getting into logistics. Amazon's making some big moves to cut carrier dependency, and drone delivery just got a $200 million vote of confidence from Zipline. Rachel: That one fires me up, and honestly, the post-purchase window is where repeat customers actually get built, and most merchants are just ghosting it, leaving money on the table. Speaker 3: The ROI math on retention versus acquisition is not even close. Repeat customer margin is where the real profit lives. Rachel: Not even a little bit. Plus, we've got the Amazon Big Spring Sale, a brand that hit a million dollars in year one with zero paid ads. Speaker 3: Which I have some questions about. Survivorship bias is real and I need to dig into the actual CAC and LTV numbers. Rachel: I knew you would. Okay, lot to get through. Let's kick it off with AI Commerce and with the ChatGPT Shopping. Shopping play actually means for merchants: Three easy ways to get your questions to us: check the description for our web form, text us at seven four seven two nine three four six one two or call the same number to leave a voice question. However you reach out, we can't wait to hear from you! OK, so I have to start with the story. Last week I'm making pasta, right, flour on my hands, can't touch my phone, and I just go, hey, order me some olive oil, and my phone just does it. Finds my brand, checks the price, done. Speaker 3: Okay, that's either amazing or terrifying. Probably terrifying if you don't own the data. And I learned the hard way, whoever controls the plumbing controls the margins. Rachel: Both, both, Rachel. But here's the thing: that moment is basically what OpenAI just made real for, like, millions of people. ChatGPT is now a shopping channel. Speaker 3: Wait, like actually transacting shopping, not just here are some links you'll never click. Rachel: Actually buy things. They've integrated Shopify storefronts directly into the chat. You browse, you buy, you never leave the conversation. Speaker 3: Hmm, okay, so how does that actually work? Rachel: So you're chatting with ChatGPT. Ask it to find, I don't know, a stand- a standing desk under 500 bucks, it pulls up products, images, prices, reviews. You hit buy, Shopify handles the checkout in the background, done. Speaker 3: Hmm. Okay, that's actually kind of wild. And I spent years optimizing product pages for humans. Now I have to optimize for a bot reading your metadata. Garbage in, garbage out applies twice as hard here. Rachel: Right, exactly. It's a full commerce layer sitting on top of the conversation. Speaker 3: And Shopify is the infrastructure behind it. Rachel: For now, yeah, but here's where it gets interesting. Shopify is now opening this up to non-Shopify brands, too. Speaker 3: Wait, seriously? So Shopify's not gatekeeping this? That tells me they want to own the plumbing underneath everyone's commerce, not just their own merchant layer. That's the real play. Rachel: Nope, which tells you everything about what they're actually after. They don't just want Shopify merchants in this network. They want to own the AI commerce layer, full stop. Speaker 3: Okay, that is a big swing. Whoever controls the commerce layer, the transaction infrastructure, controls the margins. I've seen this movie before at Amazon and Walmart. Rachel: Massive swing. And honestly, from an operator lens, it makes total sense. If AI becomes how people discover and buy things, you want to be the rails underneath all of it, not just your own merchants. Speaker 3: Right, right, so they're positioning as infrastructure, not just a merchant platform. Speaker 4: Exactly, which is a very different game. Speaker 3: Okay, so let me push on the operator side of this, because I've seen hype cycles before. What's the actual signal a merchant should watch before they put real resources behind this? Speaker 4: Good question. So the thing I keep coming back to is product data quality. Speaker 3: Yeah? Speaker 4: Like, think about how the AI picks what the surface is pulling from your Feed. Your Feed. Your titles, your descriptions, your attributes, your reviews. If that data is Messy, you're invisible. Speaker 3: Yeah, that tracks with what I've seen at scaleIve. Garbage in, Garbage out. Your Feed is your storefront now. Messy data, inconsistent titles, blurry images. You're invisible to the AI, Period. Speaker 4: Exactly. And this is where my old CRO brain goes a little haywire. Because for years we optimized product pages for humans. Good copy, nice photos, clear CTAs, now you have to optimize for a model that's reading your metadata. Speaker 3: So the shopper is a bot and I'm optimizing product metadata instead of landing page copy. That's a complete inversion from my CRO days. And honestly, I hate that I can't control the creative moment. Speaker 4: The shopper is a bot, which means your product feed is basically your new storefront. Speaker 3: Okay, I buy that. So feed quality is the foundation. But what's the real ROI signal? Like what KPI tells a merchant this channel is worth serious resources? Sources vs. Hope with a media budget attached. Rachel: So I'd say watch your assisted conversions first. If you're seeing ChatGPT show up as a touchpoint in your attribution, even in last click, that's real signal, especially early. Max: Especially when the channel is less crowded. Rachel: That makes sense. Low competition, high intent, and you're capturing share before everyone else wakes up. That's real first mover advantage. Max: Right, and the intent piece is huge. Someone asking ChatGPT to find the best protein powder under thirty dollars is not browsing. They are ready to buy. Rachel: High-intent, low friction, Checkout embedded. Honestly, that eliminates the abandonment tax I've been fighting for years. From an ops perspective, that's margin recovery at scale. Max: It really is. I mean, I spent years hunting down Checkout drop off. This basically eliminates the whole problem. Rachel: I mean, you'd think someone invented this during my checkout audit phase. Max: Right? Where was this during my checkout audit phase? Rachel: Okay, so feed quality. A intense signal, assisted attribution, those are your three KPIs. Everything else is theater, and I've approved plenty of expensive theater before. Max: Yeah, and honestly I'd add one more, review volume, because the AI is surfacing products partly based on social proof. Thin reviews, you're getting buried. Rachel: Good catch. So the whole product data stack matters. Max: The whole stack. This is the thing that I think a lot of merchants haven't fully woken up to yet. too yet. Your feed isn't a logistics thing anymore. It's a marketing asset. Rachel: And that raises the thing that actually keeps me up. If the AI is deciding what to surface and the shopper never sees your brand name at purchase, who's building the relationship here? Or are you just renting shelf space on someone else's platform? Max: Yeah, that's the part that keeps me up a little. Rachel: Because you can optimize your Feed all day, but if there's no brand moment at the transaction... production? What are you actually building? A commodity or a defensible business with margin? Max: And the data side is tricky too, Right? Right now advertisers are getting almost nothing back from these AI channels. Rachel: Which is a big problem. I've seen this pattern before. Advertisers getting almost no data back from AI channels? Blind spend with zero visibility is how you burn budget fast. So how do merchants actually get the measurement signal back? Max: That's exactly the question, and I think the answer is starting to come into focus, but it's going to surprise some people. Yeah, I've been poking around on this, and it's kind of wild. You know, for my CRO days, measurement is everything to me. If I can't measure it, I can't optimize it. That's not a channel, that's a hope. Rachel: A hope with a media budget attached. Max: Exactly. And look, early day is fine, but the question is how long brands are willing to spend into a black box. Rachel: Right, and I've been burned on this before. At scale, I've watched this at Amazon and Walmart. You need at minimum a CAC signal. Without that, you're flying blind. That's not a marketing channel, that's a bet. Max: Yeah, yeah. So what do you do? You either wait it out, or you treat it like a brand awareness. Fairness play and set expectations accordingly with your finance team Rachel: I mean, that works if your CFO is patient and your board doesn't care about unit economics. Most don't have that luxury, which brings me to something Visa published that I think is actually more interesting than people are giving it credit for. Max: Oh, the trust signal research? Rachel: Yeah, so Visa did consumer research, and the headline is, people are fine with AI doing the shopping for them, picking the product, comparing prices, all of it. Max: Right. Rachel: But the moment of purchase they still want something human legible, a recognizable logo, a checkout flow that feels bulletproof. I run a thousand checkout audits at scale. Trust signals are one of the highest leverage levers most merchants completely ignore. Max: So interesting, because it's exactly what I spent years wrestling with in my CRO days. You'd have a perfect flow, but wrong trust badges or unfamiliar payment icons? Conversion tanked. Rachel: It's muscle memory, right? Consumers have trained themselves to look for certain signals before they hand over card details. Max: And now the wild part, the shopper might be a bot, but the human still has to feel good about what the bot did on their behalf. Rachel: Which means your brand has to hold up in two places now. now, the product data layer, where the AI is evaluating you, and the checkout moment, where the human is double-checking what the bot decided for them. Max: So here's the thing. If you're a merchant with a sketchy checkout or using some no-name payment processor, that's a blocker, even if the AI loved your product feed. Rachel: Yep, and here's the practical bit. If your brand experience doesn't translate when a bot is doing the shopping, your conversion problem just got exponentially harder to fix. x. You're not optimizing for one audience anymore; you're optimizing for an algorithm and a skeptical human approving the final step. Max: You're optimizing for an audience of one algorithm and one anxious human approving the final step. Rachel: Honestly, that's the job now. Max: So clean product data, recognizable checkout trust signals, and find some way to get attribution from these AI channels, even if it's imperfect. That's the checklist. Rachel: That's the checklist. And I'd add don't wait for perfect. Perfect data. Set up assisted attribution now, even rough proxy metrics, because the channel is only getting bigger. I've seen too many operators frozen waiting for clean measurement while their competitors were already moving and capturing mindshare. Max: Yeah, build the measurement muscle before the spend gets serious. Rachel: Exactly. Okay, and speaking of what happens after someone clicks buy, there's a whole other layer of infrastructure being rebuilt right now from the ground up. Max: Oh yeah, the delivery side is getting wild. Okay, so on the complete opposite end of the funnel, Amazon is over here buying everything that moves. Rachel: Literally. Max: So they acquired a robot company called RIVR, right? Last-mile delivery robotics. And the whole point is cutting dependence on third-party carriers. Rachel: Which makes total sense from an ops standpoint. When I was running logistics at scale, the last mile was always the most expensive, most unpredictable part of the chain, and it directly killed margin. Max: Right. And from a CRO angle, delivery speed and reliability are unique. are straight-up conversion levers, like a two-day promise that actually holds up, I've tested enough landing pages to know that that beats any page tweak I could run. Rachel: One hundred percent. So what's the RIVR play specifically? Is this warehoused automation or actual out-on-the-street delivery? Because the unit economics are wildly different, and Amazon doesn't spend money on stuff that doesn't hit margin targets. Max: Out on the street, think sidewalk robots, last-mile in dense areas. Because Amazon wants robots doing the easy route so their human network handles the messy ones. Rachel: Smart allocation. Okay, what about USPS? I heard those contract talks have been rough, because Amazon's leverage game there is entirely about whether USPS needs them more than Amazon needs USPS. Max: Yeah, stalled is probably the nice word for it. So Amazon is quietly hedging; they've tightened up their relationship with FedEx. It's Expanded free returns through that network. Rachel: Wait-FedEx? I thought Amazon had basically written off FedEx a few years back? Max: Right-full circle. But it makes sense. If USPS gets more expensive or less reliable, you need a backup that's already at scale. You can't build that overnight. Rachel: And free returns through FedEx is actually a smart retention play, too, not just a logistics one. I've seen returns friction kill repeat purchase rates at scale. Lower friction, lower CAC on repeat. Max: Exactly; you lower the return barrier, you lower the purchase barrier. I've seen that correlation in CRO data enough times that it's not even debatable at this point. Rachel: Nodding. So Amazon is basically building a carrier portfolio the same way you'd diversify a supplier base: whoever controls the shipping plumbing controls the margins. They learned that lesson. Max: That's a clean way to put it. And then there's the drone piece. Rachel: Oh, Zipline? Max: Zipline raised another $200 million, which I know sounds like one of those press releases we've all been hearing for five years, but they're actually live in markets where ground delivery is slow or expensive. Rachel: I mean, I want to believe in drone delivery, but I've been hunting for the real unit economics on this for years, and they're hard to find. Max: Yeah, yeah, same. But the use case that actually works is rural or geographically tricky areas where the truck economics just don't work out, that's where drones start making sense. Right, because you're not trying to beat Amazon Prime in Seattle, Rachel: you're trying to serve a town where the nearest distribution center is two hours away, and the math on delivery margins actually works. Exactly. Max: Exactly. Niche, but real. Rachel: What does all of this mean for a merchant, though? Like, if you're not Amazon, does any of this matter directly, or is this just... Just big platform theater. Max: Well, kind of indirectly, yeah. The RIVR and FedEx moves raise the bar on delivery expectations for everyone. Once customers get used to faster, more reliable Amazon delivery, they start expecting that everywhere. Rachel: So the merchant problem is expectation creep. Max: Yep, and the practical response is, first, audit your carrier mix; don't be over reliant on one network; second, your returns experience matters more than most merchants think. Rachel: Yeah, and I'd frame returns as a marketing cost, not just an ops cost: If a smooth return brings someone back for a second purchase, the CAC math inverts: you're actually paying for retention, not shipping. Max: That's the reframe right there. And honestly, this is where most merchants miss the opportunity; the purchase comes first, obviously, but you know what they almost completely ignore? Rachel: The window right after someone buys. Max: Yep, everything that happens between the order confirmation and... actually, that's where we're going next! Okay, so shifting gears a bit, let's talk about the money merchants are straight up leaving on the table right now. Rachel: Post-purchase, yeah, this is the one that kills me. I've watched fortunes get left on the table here. Max: Right? Like the package ships and then crickets. Maybe a your order is on the way email with zero personality. Rachel: I've checked this pretty closely from the ops side. The post-purchase window is genuinely where repeat purchase behavior forms. and most merchants are just leaving margin on the table by treating it like a shipping formality. Max: It's wild! You spend so much on acquisition, you get someone to convert, and then you basically ghost them until they complain about shipping. Rachel: And then you wonder why your repeat purchase rate is flat and your CAC keeps climbing. Max: Exactly. So the 2026 Shopify data I'm seeing on this is crystal clear: personalized post-purchase communication. Amazon outperforms generic order updates by a wide margin on repeat purchase rate. Like, it's not even close. Rachel: How wide a margin are we talking? What's the actual LTV lift on personalized post-purchase flows? Max: Significant enough that merchants running personalized flows are seeing Rachel: Wow. Max: meaningfully higher second purchase rates compared to the thanks for your order, here's your tracking number crowd. Rachel: Which, when you think about it, makes total sense. You already have their attention, they just bought something, that's the highest ROI engagement moment you're ever going to get. Max: It's actually the highest intent moment you'll ever have with a customer, and most brands are sending a FedEx logo and a delivery estimate. Rachel: I mean, I'm guilty of having approved those emails before. It's embarrassing in retrospect and expensive in terms of what I could have been doing with that attention. Max: chuckling. We've all been there. Rachel: So what does good actually look like here? From an ops standpoint, I always push teams on measurable retention KPIs, second purchase rate, time to repurchase, LTV by cohort, but the execution varies wildly. Max: Yeah, so a few things that actually move the needle. First one is product education in the transactional email. Not some pushy promo, just genuinely useful content about the thing they bought. Bought, how to use it, how to get more out of it, that's the stuff that sticks. Rachel: That builds product affinity. And more importantly, it builds repeat margin. Max: Second is smart reorder timing. If someone buys a 30-day supply of something, don't email them on day two. Rachel: Right. Max: Use the purchase history, figure out when they're actually running low, and show up then. Rachel: This is where having clean data really pays off. Garbage in, garbage out. If your customer data is a mess... You can't segment, you can't time this right, and you're basically burning email on broadcast blasts. Max: 100%. And third, loyalty nudges that don't feel like spam. The bar there is just, does this feel like it was written for this specific person, or does it feel like a broadcast? Rachel: The broadcast problem is real. I've seen brands with genuinely good products crater email engagement because they're treating post-purchase as a logistics issue. statistics notification instead of a revenue channel. Max: And then they come back saying email doesn't work anymore. No, your email doesn't work. Rachel: Right, the channel's fine. Max: The channel's fine. So, I mean, the reframe here is pretty simple. Post-purchase isn't a customer service function, it's a revenue function. Rachel: That's the KPI shift I'd push on hard. If you're not measuring second purchase and time to second purchase as core metrics, You're flying blind on retention. And retention is where profitability actually lives. Max: And honestly, for established merchants, this is where the long game gets played. Acquisition costs are going down. Rachel: No, they're not. Max: So the brands that win are the ones who figured out how to make the first purchase lead somewhere. Rachel: All right, solid place to land on that. And coming up, we've got a pretty interesting contrast story. Amazon's Big Spring Sale driving real consumer traffic. The effect on one end, and a brand called Cove Essentials hitting 1M in year one with Zero-Paid-Ads on the other. Max: Yeah, two completely different playbooks, both working. That's worth getting into. All right, shifting gears before we let you go, Amazon's Big Spring Sale is live. And Rachel, the device deals are wild. Rachel: Of course they are. Amazon discounting its own hardware is never accidental. Max: Right, right. Echo, Kindle, Fire TV. When Amazon cuts prices on devices, they're not giving stuff away. They're buying your data pipeline. Rachel: Every discounted Kindle is a reading behavior feed. Every Echo... Free Echo is a purchasing signal. The sale is basically customer acquisition wrapped in a Spring theme, and Amazon owns the margin on both the device and the data. Max: Exactly. So if you're a merchant, that device penetration matters. More Alexa households means more voice-driven reorders. Keep your product listings sharp. Rachel: Okay, but let's talk about the story I found more interesting this week. Kov Essentials. Year one, a million in revenue. zero paid ads. Now here's what I want to know. What's their actual customer acquisition cost and is it repeatable? Max: Zero? Like not low spend, zero? Rachel: Zero. Organic content, community building, word of mouth. That's it. I mean, I'd want to see the attribution on that, but on the surface, yeah, no paid media spend. Max: I mean, look, coming from years in CRO, my instinct is always to test paid channels. Bills build spreadsheets, hunt the CAC. So hearing zero paid spend makes me wonder, is this repeatable or did they just catch lightning in a bottle? Rachel: That's exactly where I land. Survivorship bias is real. I've watched this at scale at Walmart and Amazon. For every Cove Essentials, there are 50 brands that went organic only and stalled at 50,000 and nobody talks about them. Max: Yeah, fair, but here's the thing. The mechanic isn't luck. Community trust converts different than paid traffic. Lower CAC, higher LTV, genuinely better unit economics. I've tracked that in the data over and over. Rachel: So the ROI case is there if the community is real, but here's the operator question: Can you actually build that intentionally with repeatable KPIs, or is it just survivorship bias and luck? Max: Probably both. Like, you can engineer the conditions but not guarantee the outcome. the outcome, which is uncomfortable for anyone who likes predictable growth levers. Rachel: Spoken like a true CRO guy. But honestly, the ones who thrive are the ones who can measure it, even the organic stuff. If you can't track it, you can't repeat it. Max: Guilty as charged. But real talk, the one thing I'd push into next week is audit your organic content, not your ads. Because whether it's an agent or an algorithm reading your brand, that's the signal they're actually pulling. Okay, that's a wrap on week 13. Rachel, honestly, this one had me thinking the whole way through. Rachel: Same. And I think if there's one thing to take away, it's this. The AI shopping shift isn't coming, it's already here. You're either measuring for it or you're getting left behind. Max: And for me, the olive oil moment really stuck. Like flour on my hands, phone in my pocket, and the whole transaction just happens. That's the world merchants are selling into now. Rachel: Right. And the trust question I kept coming back to, you know, if there's no human in that decision loop, how does your brand even show up? That's where your product feed becomes your pitch. That's what operators need to sit with. Max: Your feed is your pitch deck now, not to a buyer, to an algorithm. Rachel: I mean, honestly, that framing kind of changed how I'm thinking about this. Your data quality isn't just an operations problem anymore, it's your conversion layer. Max: Okay, so if you got something out of today, please subscribe and drop us a review. It actually helps more than you think. Rachel: And hey, tell a merchant friend. Seriously, this stuff matters at scale and most operators aren't thinking about it yet. Thanks for hanging with us. Max: See you next week.