Week 45 - ChatGPT Brand Mentions: Why Cited Pages Don't Guarantee a Name Drop, Plus Shopify's Agent Checkout, an Adobe Forecast Refresher, and New Ecommerce Tools

MaxWelcome to the Checkout Point, your quick dive into this week's ecommerce 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.

RachelOkay, quick confession. I asked a chatbot to recommend a good running shoe last night, and it named three brands I've never heard of.

Speaker 3Max, that's a you problem.

RachelMaybe, but it got me thinking about this whole episode.

Speaker 3Oh?

RachelWelcome to the Checkout Point. I'm Max, and that's Rachel, who will ask me what the ROI of my shoe shopping was.

Speaker 3Zero. Next question.

RachelSo first up, a quick refresher on Adobe's holiday forecast. We've covered it before, but there's one number inside it worth another look.

Speaker 3The AI traffic one.

RachelRight. Because then we've got new stuff: Shopify's Agent Checkout and fresh data on ChatGPT brand mentions.

Speaker 3Wait, so what's the one question running through it all?

RachelAI traffic is growing, but what's it actually worth to a seller and who gets found and named?

Speaker 3Hmm. Those might be two different answers.

RachelAnd that's the thing. A visit can look great on a dashboard and still tell you nothing about what the shopper was actually told on the way there.

Speaker 3Right. One question about the visit, one question about the recommendation. Two questions, one episode, and I'm keeping score.

RachelTotally. So anyway, let's do the refresher fast and get to the AI traffic number inside it. Okay, so Adobe's forecast release is dated September 28th. The headline is two hundred and seventy-five point one billion dollars in US online sales from November 1st through December 31st.

Speaker 3And that's up six point seven percent, which, fine, everybody expects a record. So what's the surprise?

RachelFair. The surprise is smaller. It's inside the details. But first, Cyber Monday.

Speaker 3Still the biggest day?

RachelAdobe says yes, over fifteen billion dollars. Other outlets print fifteen point one or fifteen point two, so I'm just saying over 15 and moving on.

Speaker 3Smart. Nobody wants to fight about a tenth of a billion.

RachelRight? Someone in the comments always does, though. Adobe also says Black Friday grows faster because discounts start earlier.

Speaker 3Wait, how much earlier are we talking, and how deep? Is this a polite nudge or a fire sale?

RachelAdobe says it expects discounts of up to thirty percent during Cyber Week. Up to, so that's the ceiling, not the average.

Speaker 3Earlier discounts, so margin pressure shows up sooner. I've seen that movie before.

RachelI bet you have. Okay, the number I actually care about.

Speaker 3Go.

RachelDigital Commerce 360 covered the forecast on September 29th. It reports Adobe projects AI traffic to US retail sites up one hundred and thirty percent year-over-year.

Speaker 3Wait, one thirty? Like, more than double?

RachelMore than double. And the biggest jumps land on Thanksgiving, Black Friday, and Cyber Monday.

Speaker 3Does that same coverage put a dollar figure on Black Friday itself?

RachelIt does. Twelve point nine billion dollars, up nine point two percent. So Black Friday growing faster than the overall forecast isn't just a talking point, it's in the numbers.

Speaker 3Hmm. So the exact days when everybody's already slammed.

RachelExactly the days. That's wild, right? A brand-new kind of visitor shows up on the busiest days of the year.

Speaker 3Wild, sure, but busiest also means stretched inventory, stretched support, and stretched attention.

RachelThe same coverage also lists Cyber 5 at forty-seven point five billion dollars and electronics as the top category at sixty-three point three billion dollars.

Speaker 3Electronics. Big-ticket stuff. I'd love to know whether the AI visitors even land there.

RachelGood question, and I don't have that breakdown, so I'm not gonna guess.

Speaker 3Okay, but traffic is traffic. I can buy traffic. Plenty of it is junk.

RachelTotally.

Speaker 3So if AI traffic is climbing that fast, what is it worth to a seller? Because a visit isn't a sale, and I wanna know the number that pays rent.

RachelI mean, come on, that's the whole question, and I'm not touching it yet.

Speaker 3Convenient.

RachelHey, I need something to dangle. Speaking of AI traffic, Shopify just put some numbers on what those visits look like.

Speaker 3Oh, now I'm listening.

RachelOkay. So Shopify, a merchant-focused October rundown from a site called Fudge says Shop MCP landed October 13th.

Speaker 3Remind me what MCP does for a normal seller.

RachelAny agent that supports MCP can search a store's catalog, build a cart, and check out. Shop Pay handles it for signed-in shoppers.

Speaker 3And the partnership part?

RachelThere isn't one. No deal required , which is the opposite of how most retail integrations go. You know what I mean?

Speaker 3Wait, really? When I was at Walmart, getting into anybody's pipeline meant six months of legal.

RachelRight? Now, an agent just walks in the front door.

Speaker 3Though walking in the front door cuts both ways. A door that's open for agents is open whether or not your catalog is ready.

RachelGood point. The rundown gives us the open door, not a readiness checklist, so the homework lands on the merchant.

Speaker 3Okay, but translate that for a seller. If I run a store on Shopify, what actually changes on Monday morning?

RachelGoing by the rundown, your catalog is something an agent can read, and a card is something it can build. What the rundown doesn't tell us is how many agents are actually knocking, so I wouldn't assume a flood of traffic.

Speaker 3Mm. But that rundown is a secondary summary. It tells merchants to confirm details in Shopify's changelog.

RachelFair. So treat it as a pointer, not gospel.

Speaker 3Exactly. Because pointers are cheap, but a checkout flow that misprices your catalog is not.

RachelWhich is a fair point. If an agent reads your catalog, the catalog is your storefront. Titles, prices, availability, all of it has to be clean.

Speaker 3A messy product listing used to be an annoyance, now it's something a machine reads and acts on, and I want those details right before somebody's cart depends on them.

RachelSo if I were a merchant, the walkthrough is boring but useful. Open your top product pages and read them like a stranger would. Is the title clear? Is the price right? Does it say what's in stock?

Speaker 3And do it before the season, not during. Nobody wants to discover a pricing typo with a promotion already running.

RachelSpeaking of prices, the same rundown says agents have seen compare at prices since October 9th.

Speaker 3So the strikethrough price is visible to the bot.

RachelWhich means your discount framing is readable by machines, not just humans.

Speaker 3So the sale price and the original price both matter. If my compare at price is sloppy, the bot sees that too, right?

RachelYep, and it says Google's Universal Cart went live in the US on October 15th.

Speaker 3Two days after Shopify opened up. That's tight.

RachelTight timing, sure. I'm not claiming they're connected. The rundown just lists the both.

Speaker 3Good. Don't hand me a causal story off a calendar. Okay, but now I want the number I asked about last segment.

RachelWhat AI traffic is worth. Here's where it turns. A weekly agentic commerce newsletter reports a Shopify finding from October 13th.

Speaker 3Go on.

RachelShopify says AI-referred sessions converted eighty percent better than organic search.

Speaker 3Eighty. Say that again?

RachelEighty percent better conversion, and average order value ran fourteen percent higher.

Speaker 3Fourteen percent on order value matters, too. Better conversion with smaller baskets would be a different story.

RachelRight. So both moved in the good direction as reported. Compared to organic search, more of those sessions converted, and the average order was bigger.

Speaker 3That's wild. Okay, I'm interested, but what's the actual ROI?

RachelI knew that was coming.

Speaker 3Because conversion against organic search is one comparison. Who's in each group? How big is the AI group?

RachelThe newsletter doesn't give that, and it's reporting Shopify secondhand, so I'm attributing the figures to Shopify and leaving it there.

Speaker 3That's my problem with any percentage. A big relative lift on a small group is still a small group.

RachelAnd that's the CRO nerd in me waking up. Picture a store where only a modest slice of visitors ever convert. A huge relative lift on that is great news, but it might be a handful of extra orders. I'm making up the shape there to make a point, not quoting anything. The lift is what was reported. The size of the group is what we don't know.

Speaker 3Fine. Then the caveat matters more.

RachelIt does. The newsletter stresses these figures measure AI-referred visits.

Speaker 3Visits.

RachelVisits, not purchases an agent completed on its own.

Speaker 3So a person asked a chatbot, clicked a link, landed on the site, and then a human did the buying.

RachelProbably the common picture, but the newsletter only says what the figures measure. It doesn't say what any one shopper did.

Speaker 3Okay. Careful. Still, that's a different animal than a bot with a credit card.

RachelAnd the wording matters for how you talk about it internally. A visit from an AI tool is somebody who is curious enough to click. An agent completing a purchase is a different kind of event, and the numbers are about the first kind.

Speaker 3So the careful phrase is AI referred, never agent bought. Otherwise, someone builds a plan around a channel that hasn't shown up in the data yet.

RachelTotally. And the newsletter cites consumer research that points the same way. Most shoppers still check AI recommendations.

Speaker 3Of course they do. I double-check my own phone's suggestions on where to eat.

RachelAnd they hesitate to delegate the purchase itself.

Speaker 3So the plumbing is open, and the shoppers aren't walking through it yet.

RachelMm-hmm.

Speaker 3Here's another bottleneck. The same newsletter says banks have moved to restrict agent access to customer accounts.

RachelWait, really? Right as Shopify opens the door?

Speaker 3That's what it reports, and that's the part I'd put on a whiteboard. Checkout is one step. Somebody still has to move the money.

RachelSo the pipe at one end is wide open, and at the other end, somebody's tightening a valve.

Speaker 3Which I'd call a bottleneck, and I don't love it when merchants plan around the wide end only.

RachelThere it is. Supply chain Rachel.

Speaker 3Somebody has to be.

RachelThe practical read on the bank part is that a merchant can't control that end. What you can control is how clean your product data is.

Speaker 3Control what you can, watch what you can't. That's most of operations, honestly.

RachelOkay, so what did we learn on the question you left hanging? AI-referred visits look valuable per Shopify's numbers.

Speaker 3Valuable visits, not proof that agents are buying alone.

RachelRight. Those are two separate claims, and only one of them has a number on it.

Speaker 3And I'd say read the eighty and the fourteen as a reason to look closer, not as a number to drop into a forecast.

RachelLook closer first, promise later. I like it.

Speaker 3So the next question is upstream. If AI is sending people to stores, which stores is it even sending them to?

RachelOh, that reminds me. There's fresh data on which brands ChatGPT actually names.

Speaker 3Names, not just links to.

RachelHold that thought. Okay, so fresh off the press, Search Engine Journal ran a piece on October twenty-sixth, and the headline alone made me wince.

Speaker 3Go on.

RachelThe gist is that some brands sitting on ChatGPT's cited pages still got very few mentions. Being cited and being named turn out to be two different things.

Speaker 3Wait, say that again. Cited but not named?

RachelYep. The page shows up as a source. The brand on it doesn't show up in the answer.

Speaker 3Huh. So the footnote gets credit, and the brand gets nothing.

RachelBasically the backup dancer who never gets a name on the poster.

Speaker 3Okay, but what's the pattern? Who gets skipped?

RachelSo Search Engine Journal says the new comparison prompt data found something sneaky. Brands that logged out ChatGPT rarely named when it didn't search also showed up less when it did search.

Speaker 3Even with the cited page coverage?

RachelEven with the cited page coverage, search didn't rescue them.

Speaker 3Hmm. That's the part I didn't expect. I figured search was the equalizer.

RachelMe too. I'd have bet the money on it.

Speaker 3And now I have to ask, what does the piece say about how the data was gathered?

RachelI'm only going with what the summary gives us. It describes comparison prompts, logged out ChatGPT with and without search. Beyond that, I'm not gonna guess at the setup.

Speaker 3Fair. I'd rather you stop there than make something up.

RachelTotally. And one more caution: this is a secondary writeup, so treat it as a signal, not a verdict.

Speaker 3Okay, so what does this do to the return on all the effort people put into getting cited?

RachelThere she is.

Speaker 3I'm serious. If I'm a seller, I've spent time and money getting my pages cited. Where does that land on the P&L if the answer never says my name?

RachelRight, and that's the cost with no receipt. You did the work, the source list has you, and the shopper reads a recommendation that skips your brand entirely.

Speaker 3So being cited is a line item, not a result.

RachelWhich sounds like a great title for a very sad conference talk.

Speaker 3Keynote, 10:00 AM. Bring tissues.

RachelBut seriously, it answers what we were chewing on earlier. Getting your pages into the pile doesn't guarantee you show up in what the shopper reads.

Speaker 3Mm. So the question to ask isn't whether you're cited, it's whether the brand name is in the answer.

RachelExactly. The data says one doesn't buy you the other. Okay, before I forget, we've got a handful of tools and quick hits for the week.

Speaker 3Oh, good, something I can actually put on a spreadsheet.

RachelOkay, tools time. Practical Ecommerce put out its weekly new tools roundup dated October twenty-first, and three items jumped out.

Speaker 3Only three? I've seen that list run long.

RachelThree that matter for us. First, Algolia added native Shopify B2B search.

Speaker 3Native, meaning no duct tape.

RachelThe roundups frames it as B2B search for Shopify, so if you sell to other businesses, that's the one to look at. If you only sell to consumers, you can probably skip it.

Speaker 3And my test for any search tool is simple. Does it help a buyer find the thing, or does it just add a line to the invoice?

RachelRight. Second, Lightspeed launched agentic capabilities. They recommend actions and prepare them, and the retailer approves.

Speaker 3Hold on. The human still clicks yes?

RachelThe human still clicks yes.

Speaker 3That's the part I care about. A tool that recommends is a suggestion. A tool that acts without sign-off is a liability.

RachelAnd the roundup's wording is recommend and prepare, then the retailer approves. So the sign-off is built into the pitch.

Speaker 3Good. I've watched software prepare things nobody approved, and unwinding it was no fun.

RachelOf course you have.

Speaker 3Dry wit, real scar tissue. What's the third?

RachelMarklo raised five million dollars in seed funding. It's an agentic marketing platform for Shopify brands.

Speaker 3Five million. Seed. So it's early.

RachelVery early. Tool to watch, not tool to bet the quarter on.

Speaker 3Notice the theme though. Lightspeed and Marklo both have agentic in the pitch.

RachelSame word we kept hearing in the Shopify segment. Agents on the shopper side, agents on the seller side.

Speaker 3But what does it actually return?

RachelRachel, it's a seed round. The return is a slide deck.

Speaker 3Fair. Watching, not buying.

RachelAnd for the record, five million dollars in seed funding says investors see something. It doesn't say merchants have proof.

Speaker 3Right. Funding is a vote of confidence from investors, not a result in your store. Different columns on the spreadsheet.

RachelTotally. Okay, one that's more of a calendar thing for Shopify merchants.

Speaker 3Uh-oh. Calendar things are where money quietly leaks.

RachelA Shopify rundown from Fudge, a merchant-focused blog, says no new storefront script tags have been allowed since October first.

Speaker 3Since October first? So that's already in effect.

RachelYep, and the same rundown says the existing ones get removed March seventh, twenty twenty-seven.

Speaker 3Wait, removed, as in gone?

RachelThat's how the rundown puts it.

Speaker 3So if some old widget on your storefront runs on a script tag?

RachelIt could just stop working. Maybe your reviews badge, maybe a pop-up, whatever you bolted on years ago.

Speaker 3That's a supply chain problem with a date on it. Somebody has to audit every tag before March.

RachelSo how do you even start? List every app and widget on your storefront and ask whether each one depends on a script tag. Boring, but it's the whole job. Once it's listed, sort by what would hurt most if it vanished. Something near your checkout or your product pages matters more than a decorative badge.

Speaker 3Right, and ask each vendor what their plan is. That's a question, not a claim. The answer is theirs to give, and you want it in writing before anyone is scrambling.

RachelAnd it's a secondary summary, so check Shopify's own change log before you panic or tell your developer.

Speaker 3Mm, developer gets a ticket today, not in February.

RachelAnd put both dates in the ticket. October first for the block on new tags and March first, twenty twenty-seven for removal, so nobody has to go hunting later.

Speaker 3Plus a spare week before March because audits always find one more widget nobody remembers adding.

RachelHa. Right. Nobody's ever been early to a deadline.

Speaker 3I'm early. I'm always early.

RachelYou're a spreadsheet with legs.

Speaker 3Thank you. Okay, quick hit from me. A News-Pressnow.com article from October twenty-sixth reports on a joint study by WDL and EBANX.

RachelNever heard of the pairing. What'd they find?

Speaker 3Emerging markets will generate sixty percent of new online spending over the next decade.

RachelSixty? Wait, really?

Speaker 3And most of it comes from consumers who don't have a credit card.

RachelLet me restate that so nobody walks away with the wrong number. Sixty percent of new online spending over the next decade from emerging markets, not sixty percent of all online spending.

Speaker 3Right. New spending, which is the growth slice. The headline also mentions Asia adding more than seven hundred million middle-class adults by twenty thirty-six. I only have the summary, so I'm not gonna break it down by country or payment type.

RachelHuh, that one hits different after the agent checkout talk. A bot with... Okay, I'm not reusing that

Speaker 3You were so close.

RachelBut seriously, if the next wave of buyers isn't on cards, what does checkout look like for them?

Speaker 3That's the question. I don't have the answer from the article, and I won't invent one. But if you sell cross-border, ask your payments provider what they offer beyond cards.

RachelSimple and cheap to ask.

Speaker 3Cheapest research there is. Okay, last one is yours.

RachelOne line. Search Engine Journal ran a piece from Brent Csutoras on Reddit CEO Steve Huffman at Mod World.

Speaker 3Reddit again. Brands love to show up there uninvited.

RachelThe piece says Huffman's vision points Reddit towards deeper communities.

Speaker 3And?

RachelAnd that gives brands a stronger reason to build subreddits of their own.

Speaker 3Hmm. Owned the room instead of crashing someone else's.

RachelOwned space does mean owned work, though. A subreddit of your own is a community to tend, not a billboard.

Speaker 3And community needs a person with time, not a campaign that ends in January. I'd ask what it costs per month before I'd ask what it returns.

RachelThat's the pitch. Whether it pays off, no idea. The piece is a take, not a case study.

Speaker 3So a maybe with a moderator's badge.

RachelI'll take it. So anyway, I think that's the week.

Speaker 3Then let's wrap this up.

RachelOkay, quick landing. AI-referred visits are a real signal. They are not proof that an agent bought anything.

Speaker 3Right, and being cited on a page doesn't mean the model says your name out loud.

RachelTwo takeaways, zero tissues required.

Speaker 3Max, you're not allowed to reuse the tissue joke. Anyway, my action item is boring, which is why it works. Pull your AI-referred sessions into their own line in your reporting. Then ask what each one actually costs you and what it actually returns.

RachelSo, like, tag it before you celebrate it.

Speaker 3Exactly. A visit you can't price is a hobby.

RachelOuch. My side project just felt that.

Speaker 3It should.

RachelOkay, my turn to close. If a chatbot ever cites me and doesn't name me, I'm treating it like a coworker who reads my report and presents it as theirs. Follow the show wherever you listen, so at least somebody credits us.

Speaker 3Name included. See you next week.