Week 12 - AI rewrites ecommerce playbooks: Shopify’s valuation, payments shakeups, and global growth hotspots

Max: Welcome to the Checkout Point, 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: Okay, so picture this. I'm cooking dinner, literally have flour all over my hands, and I just say to my phone, Order olive oil. Boom. The AI finds my favorite brand, compares prices across retailers, checks reviews, and completes the whole transaction without any apps, without friction, without me thinking about it once. Speaker 3: Which sounds great until the invoice hits and you realize the helpful AI just killed your contribution margin. I've seen this movie at scale. Margin evaporates before anyone notices. Rachel: Exactly. So here's the real question for the next 90 days. Where are you actually betting? Creative AI tools? Chatbots doing support deflection? Or merchandising and bid agents that are touching your ROAS and contribution margin? Speaker 3: And we're anchoring it in hard KPIs-incremental revenue, support deflection, true ROAS, contribution margin. Plus the red flags that scream expensive theater. Onboarding takes three months, metrics are fuzzy, deck says transformative instead of payback. Rachel: Right. Like if the onboarding takes three months and you still can't articulate what metric it actually owns or how payback works, I'm running. I've seen that pattern too many times in my CRO days. Speaker 3: If the vendor can't tell you payback period or owns zero accountability for actual margin impact, that's when I check out. Rachel: Then we're pivoting to Shopify. They just crushed their quarter, but the AI narrative has public markets absolutely freaking out, which honestly is kind of hilarious when you remember operators care way more about actual margin than they do about Wall Street's multiple anxiety. Speaker 3: Yeah, so we'll break down operator reality, when you actually double down on the Shopify AI stack versus when platform risk, logistics complexity, or better unit economics say test alternatives instead. I learned that at Amazon, whoever controls the pipes controls the margins. Rachel: Which sets up the fun part, the Checkout Wars sequel. Stripe, PayPal, Klarna, even stablecoin rails. Speaker 3: We'll give you a no BS payments playbook, auth rate, blended take rate, LTV by tender type, and the hard truth when testing a new rail actually lifts margin and cash flow instead of just looking cool on a pitch deck. Rachel: And then, global expansion. Duties, taxes, localization, FX complexity. This is the stuff that used to legitimately make me want to pull my hair out back in my CRO days. I'm not even exaggerating. Speaker 3: We'll map where the next billion in GMV actually lives, plus build you a simple expansion ladder and three non-negotiable metrics – CAC, repeat rate, and fully loaded logistics cost. AI as a multiplier, not a magic wand. That's the only story that matters. Rachel: All right, let's jump in. 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. Picture this: I'm in my kitchen. Hands absolutely covered in flour, phone on the counter. Speaker 3: Yeah, this is chaos-but it's exactly where the incentives flip for operators like me: no algorithm tax, no bidding wars just to get eyeballs. The AI is the funnel, and that's where margins actually hide. Rachel: I just mumble, "Hey, I'm out of olive oil. Order my usual. And that's it. No app to open, no search bar, no cart to add to. The AI pulls my favorite brand, compares prices, checks reviews, applies the right card, and Boom, done. Zero friction. Speaker 3: Right. That's when I stop nodding and pull out a spreadsheet. Cool story. What's the actual delta on my blended CAC and margin? Rachel: Exactly. I've been preaching feed the algorithm your best stuff since my CRO days. Best images, best copy, cleanest data. Speaker 3: Mm-hmm. Rachel: Now, the algorithm is your shopper. So the real question becomes, where do you actually invest first? Speaker 3: Yeah, let's get concrete. In the next 90 days, if you're mid-market, you're drowning in three shiny objects – AI creative, AI chat, and AI agents managing bids. The question is which one actually moves your P&L. Rachel: The play is collapsing the creative-test-launch cycle down from weeks down to days. Back in my CRO trenches, you'd brief a photographer, wait for edits, Speaker 3: Right. Rachel: then brief the media team, and you're already three weeks deep before you've tested anything. Now you can spin 20 product angles, test hooks, swap backgrounds all in one afternoon. Speaker 3: And that's not just pretty pictures. That's CPMs, CPCs, and CTRs moving. Rachel: Exactly. You can run real structured tests. Same audience, same budget, five AI-generated variations, kill losers in 48 hours. That's the CRO dream I've been chasing for years. Speaker 3: I like that because I can actually see the KPI. If your blended CAC doesn't drop or your click-to-add-to-cart doesn't improve in four weeks, the tool isn't earning its seat at my table. Rachel: Totally. If the vendor can't connect it to a lift in ROAS or conversion rate, it's theater. Full stop. Speaker 3: Yeah, but what about the bots people are hemorrhaging cash on? We launched an AI concierge and increased engagement time by 40%. Cool. What actually moved on revenue per session? That's when people go quiet. Rachel: Let's actually separate vibes from value here. A useful chatbot does three concrete things. Raises AOV, deflects support tickets, and saves sessions that are about to bounce. Speaker 3: Walk that through. Rachel: For AOV, think, I'm buying running shoes. A good bot says, cool, here's socks that prevent blisters, here's a hydration belt. If your attach rates or bundle adoption doesn't move, it's not working. Speaker 3: So, merchandiser brain in a bot. Rachel: Exactly. Support deflection. Max: means it answers, where's my order, does this fit, what's your return policy without a human touching it. You measure deflection rate and average handle time. And conversion? That's the save the cart moment. Someones hesitating, the bot clarifies sizing, recommends the right variant, maybe nudges with a review snippet. If sessions with bot interaction don't convert better than sessions without, you've got very expensive theater wearing an AI hat. Rachel: Very expensive. I've watched this movie before at scale. Beautiful engagement metrics, zero revenue lift, and a six-figure tab. Max: Very expensive. Rachel: Red flags I look for? Max: Wow. Rachel: Onboarding takes three months of consultants and they can't tell me with specificity what metric will move by what magnitude. Or they lead with engagement time instead of revenue per session or deflection rate. That's when I walk. Max: Yeah. If people chatted for seven minutes, is your headline? That's a customer service nightmare, not a win. Rachel: Okay, third bucket, AI agents managing bids, budgets, merchandising. This is where operators either protect margins or accidentally blow them up. Max: This is where the big money is hiding. We're shifting from campaign-led growth (humans building calendars, flighting budgets manually) to always-on agents that tune bids, swap creatives, and reorder product blocks based on real-time behavior. That's fundamentally different. Rachel: So instead of humans manually planning holiday campaigns, the system's just... Constantly rebalancing in real time? Inventory, margin, traffic, all based on live data? Max: Exactly. Think of it as a 24-7 trader plus merchandiser. It looks at margin, inventory, and click velocity and says, we're pushing product B today, not A, because A is about to stock out and B has better margin. Rachel: I love that in theory, but in practice, Max, where does a mid-market operator actually start without blowing up their margins or operations? Max: I'd start narrow and clean. One channel, one objective. For example, let an AI agent manage Google Shopping bids for your top 200 SKUs with a hard guardrail like target 20% blended margin. No ambiguity, no wiggle room. And if you're not seeing improvement in 60 days, turn it off immediately. You should see either better ROAS at same spend or same ROAS at lower spend. If your main success metric is increased impressions, that's a red flag. You're gaming the vendor instead of optimizing for actual profit. Rachel: Exactly. Impressions are vanity. I need ROAS, margin contribution, or efficiency per dollar spent. That's what moves P&L. That's what I look at. Okay, force the debate: brand doing twenty five to seventy five million dollars online. Next ninety days they can only go hard on one thing: AI creative, AI chat, AI agents. Where do you send them first? Max: So let's just ask the hard question. Is Shopify's multiple actually earned or are we just pricing in AI upside that won't hit the P&L for three years while margins stay unclear? Yeah, we're going straight there. I like it. Rachel: From a P&L brain—and this is where I get annoying—here's what I see: record GMV, analyst upgrades, buybacks, and yet the stock whipsaws because nobody can actually model where steady state margins land. That gap between story and math-that's everything. Max: Right, so you've got best quarter ever on paper, solid metrics, all the wins, but then the stock whipsaws because nobody can actually model where margins land. That gap between narrative and reality is everything. Rachel: Exactly. So if operations are solid, why are smart money asking, what if Shopify's just plumbing while Apple or Amazon own the AI experience? Max: Couple things. One, AI narrative risk. If agents like the olive oil bot I told you about flower on my hands just ordering while cooking actually start handling the cart, does Shopify stay the toll booth or does it become plumbing? Rachel: So rails versus commodity? Max: Yeah, if Shopify is genuinely the AI commerce OS that earns a premium multiple, if it's just a cheaper cart sitting behind someone else's AI magic, your margins get crushed and investors know it. Rachel: And that's the real panic. What if the AI layer lives at the platform level, Apple, Google, Amazon, and Shopify is relegated to commodity processing in the background? That's not a premium multiple. That's a toll booth. Max: Exactly; plus AI costs hit the P&L before AI revenue shows up. So investors see opex up, margins pressured, and they go, do I really want to pay ten, twelve times revenue for this? Rachel: Big R&D spend, margin pressure up front, and... here's the kicker... completely unclear steady-state margins. When I ran this at Walmart scale, that uncertainty alone would tank your capital allocation. Max: But here's where I'm actually more bullish. At the operator level, and I've been in the trenches with this doing CRO, Shopify is becoming the easiest way to plug into the future agent world. You get Shop Pay, the app ecosystem, the APIs, the AI features, all without rebuilding from scratch. Rachel: Exactly. So for a 40 million or 400 million brand, the real question isn't, is the valuation fair? I don't care what the market thinks. It's, does my business actually get smarter and more profitable using Shopify's full stack, or am I just hoping? Max: Yep. Operator reality beats investor panic every time. Rachel: I'm running a 40 million DTC brand. Should I go all in on Shopify's AI roadmap or stress test alternatives? Max: Pull out a spreadsheet. Question one, where is your margin actually made? On your own site or across wholesale and marketplaces? Rachel: That's the core insight. If your dot-com generates 70% of profit, your job is ruthless. Maximize that funnel and measure it. You're not playing investor roulette. You're protecting margin. Max: Exactly. Question two, do you have any real reason to abandon Shopify's ecosystem? Something concrete like technical limits or a blended fee stack that's actually higher than alternatives? Exactly. Spicy threads don't move your EBITDA. Operator reality does. Rachel: Yeah. Spicy threads do not move your EBITDA. Max: When is Shopify still the obvious choice? Rachel: Three cases. One, you're sub 100 million GMV and want speed. You rarely win by over-architecting at that stage. Shopify Plus apps plus AI tools is the path of least resistance. Max: Two, you actually care about conversion and CRO wins more than having some ultra custom experience. Their checkout, Shop Pay, pre-built flows. From a conversion standpoint, the baseline is genuinely solid and I've seen teams waste six figures trying to beat it. Rachel: I've watched teams waste six figures recreating Shopify's checkout from scratch. That's what I call expensive arrogance, and it kills margin before it kills ego. Three, you want to ride their AI roadmap instead of hiring your own ML team. If your goal is better recommendations, smarter merchandising, agents that handle ops, you're probably getting better ROI leaning into what they're shipping. Max: I agree. But here's my caveat. If your business is so concentrated on Shopify that one policy change nukes your margin, you need to war game alternatives now, not panic plan later. Rachel: Yeah, that's fair. Max: I'd stress test three hard scenarios: app fees spike thirty percent; they deprioritize your use case or their payments economics shift. Do you have a real backup wargame or just wishful thinking? Rachel: So default yes on Shopify, but keep a real contingency plan. Max: Exactly; and reframe the hedges; use marketplaces and retail, not as Shopify insurance, but as insulation against any single channel risk. Diversify your demand, not just your tech stack. Rachel: Right. Diversify your demand, not just your tech. Max: When should someone lean harder into Shopify's AI right now? Rachel: One, your own site is your main profit engine. Two, your team is small and drowning. Three, you've got clean data. Solid catalog, good images, tags that actually make sense. So AI can help. Green light. Max: And when should they stress test alternatives? Rachel: When your dev team is constantly hacking around platform limits, your fee stack is bloated, or your brand has complex multi-region, multi-catalog needs that Shopify doesn't handle cleanly, that pain from my CRO days is real. Max: Plus, if you're big enough that 20 to 30 basis points on payment or infracosts justifies engineering time, that ROI should be modeled and measurable, not assumed. Rachel: Totally—and one more nuance: hedging doesn't mean ripping Shopify out. It might mean decoupling some services like search, recommendation, or content so you're less locked in. Max: Yeah, Shopify is your home base, that makes sense. But don't let one policy change crater your whole PL. That's operator risk, and it keeps me up at night. Rachel: Meanwhile public markets can freak out about multiples all day, but your actual job is simple: does this stack help you sell more, at better margin, with less operational headache and pain? Max: That's the line right there. Rachel: Put it on a mug. Max: Alright, and speaking of where margin actually lives, where it's made or lost, Next, payments. Rachel: Yep. Stripe's record volumes, Klarna usage, maybe a wild Stripe, PayPal, what if? Max: And, more importantly, what that actually cost you in blended fees, approval rates, and cash velocity, because those numbers are where the real margin story lives. Rachel: Grab your virtual calculator and maybe a spreadsheet. We're diving into checkout rails next. Yeah, welcome to the Checkout Wars sequel. We talked platforms last segment, now it's who's actually eating at the margin buffet. And spoiler alert, it's almost never the merchant. Max: Yeah, because from an operator lens, and I mean this, I'm always asking, is this helping my bottom line or just padding someone else's margin stack? Rachel: Exactly. So zooming out, Stripe's pushing record volumes, Klarna's bragging about daily active users like they're a social app, and people keep whispering about a Stripe PayPal combo that would be absolutely wild. Max: Which would be wild, because then you've basically got one aggregator logo, cards, maybe one BNPL option. That's not competition, that's the merchant getting squeezed from every side. Rachel: Historically? Card networks, Visa, Mastercard, they take their cut on basically every single rail. Processors like Stripe and PayPal win on software margin. Merchants are sitting there like, cool, I got higher conversion, but my blended take rate just quietly crept from 2.5 to 3.1. And honestly, nobody's even measuring it. Max: Right. Operators don't measure the creep. They see top line sales and miss that their effective take rate. Climbed 60 basis points in plain sight. I want to see the formula monthly: Total fees divided by processed volume. That's the tax. That's real. Rachel: Yeah, net of disputes, FX, all the junk fees. That's your real tax on revenue. Max: So let's talk stablecoins and digital banks. Everyone hears USDC and thinks crypto casino and fair. But operationally, the real question is, can this meaningfully lower my true cost to process or is it just moving the fee around? Rachel: You can literally pull cohort performance by tender type in most data stacks now; if Klarna users churn faster, their higher conversion might not be worth the extra fee. Max: When does it make sense to test alt rails, stablecoins, more aggressive BNPL? Rachel: I'd say three triggers. One, your international share is climbing and FX plus fees are actually crushing your margin. Two, your auth rate is materially below what peers are seeing, like sub 89% when others are hitting 92, 94%. Three, cash flow is tight and settlement speed could genuinely move your inventory turns and working capital. Max: I'd add customer demand as a signal, not a reason. If your crypto or gaming niche hammers support asking why can't I pay with X—that's a data point to test, not to replatform. Rachel: Fair point. But I'd cap the experiments hard. Don't turn checkout into a NASCAR jersey. Pick one alt rail or BNPL partner to test per quarter, max. I learned the hard way in my CRO days: too many variables and you literally can't see what's Max: Actually working. Rachel: And this is where operational overhead kills your margin. Every new rail adds fraud surface, refund complexity, support tickets. One wrong move, and your conversion lift gets buried under operational debt you didn't budget for. Speaker 3: How do you structure a test? Rachel: Sixty to ninety days, hard metrics only. Example: BNPL test has to lift checkout conversion three to five percent with zero increase beyond fifty bps net fee and no LTV degradation. Miss it, you kill it, no mercy. Speaker 4: Same for a digital bank cross-border setup: you want at least 20 to 30 bips fee improvement or better off in target markets plus cleaner settlement timings. Rachel: And customer experience metrics matter too. Refund cycles, dispute time to resolution. Faster settlement for you is expensive theater if refunds take three weeks and crater your NPS. Speaker 4: Mm-hmm. So tying back to Shopify from last segment, most folks are totally fine staying on native payments plus one or two strategic add-ons. But the second you chase cross-border or new rails, you're playing a completely different game. Rachel: Which sets up where we're going next. Speaker 4: GLOBAL EXPANSION Rachel: Because once you cross borders, all of this-fees, FX, auth, settlement timing-doesn't just get harder, it multiplies, compounds. Most teams aren't mentally prepared for that complexity. Speaker 4: And we'll get into which corridors actually deserve that pain-India, Southeast Asia, China-EU. Rachel: Plus, you have to read whether early numbers signal a real market or just a money pit you're about to throw infrastructure at. I've seen too many teams over-invest in corridors that never hit payback. Speaker 4: Skips okay? Rachel: Okay, we just spent a segment on payments getting messy at scale, but I need the real answer: are US and EU brands operationally READY to chase growth in India, Southeast Asia, or are they just chasing the narrative? Speaker 4: Emotionally, no way. On paper, absolutely, they should be. The Netcore data is pretty clear. The next big GMV spikes aren't coming from squeezing another 5% out of the U.S. market. They're coming from India, Indonesia, Thailand that hold China-EU corridor. Rachel: Right. Speaker 4: And the wild part is the agentic shift they talk about. People there are skipping old steps. They're going straight to AI-driven discovery, chat-based shopping, super app-style journeys. Rachel: Mm hmm. Yeah, and this is the operator disconnect I saw constantly. Market comfort with a channel and your ability to actually deliver profitably at scale, two totally different problems. Speaker 4: Exactly. So if you're a U.S. brand, the question stops being, "Is their demand?" and starts being, "Am I actually willing to play their game instead of forcing them to play mine? Rachel: Exactly. Operational readiness isn't sexy, but it's where margins live or die. At Walmart and Amazon scale, demand was never the constraint. The constraint was, can I actually get this there fast, legally, and without the logistics bill eating my entire margin? Speaker 4: Mm, the fun stuff. Rachel: You've got DHL-JD corridors, customs delays, duties, de minimis thresholds moving every quarter, one tariff rule shift and your landed cost model explodes. I've literally watched well planned expansion plans crater margins in real time. Speaker 4: Yeah, those duties and tax calculators made me want to pull my hair out. You'd nail conversion rate and UX, then boom, surprise 30 bucks in duties at checkout, and the whole thing nukes. Rachel: Exactly. And here's what gives me hope. The logistics infrastructure genuinely shifted. Consolidated 3PL, regional 3PL networks, better cross-border programs, plus the AI layer actually helps if your data backbone is solid to begin with. begin with. Speaker 4: Spell that out. Rachel: Rule engines pulling real-time Duties, agents picking the cheapest compliant lane, localizing delivery promises, rewriting content for Thai or Bahasa, but, and this is critical, only works if your product feed and inventory data aren't trashed to begin with. Speaker 4: And this is where I get annoying, but I mean it. AI doesn't fix a bad ops backbone. Never has. Rachel: Totally. Speaker 4: If your inventory accuracy is trashed domestically or your 3PL can't hit 90% on time, shipping to Jakarta is just exporting your chaos at scale. You're not optimizing, you're spreading the damage. Rachel: Chilling. Now with international chaos. Speaker 4: So let's actually give people a ladder. Not every brand should wake up tomorrow and go all in on Bangkok overnight. Rachel: Yeah, let's keep it simple. And I mean simple. Test one or two regions max, not six. I've seen that movie too many times. Speaker 4: Step one, signal check. You should be thinking global when you've got solid product market fit at home, say 10 to 20 million plus online revenue, profitable or close to it, and you're seeing actual organic demand from abroad. Rachel: That's the key signal. If you've got people in India or the Philippines willing to pay insane shipping just to buy from you, they're voting with their wallets. That's your real pull signal, not a hunch. Speaker 4: Exactly. Step two, sandbox one or two regions only. If you're US-based, start with UK or EU, no culture shock. Then GCC or Singapore, high GMV, logistics friendly, low chaos. Then maybe Southeast Asia or India, not the other way around. Rachel: I'd push back harder on that. Don't even think about India first. It's huge, brutally price-sensitive, and operationally messier than most teams can handle coming in cold. That's where I've watched cash burn. Speaker 4: Right. Rachel: Two, 90-day repurchase rate. If it's materially worse than your baseline, you've got a positioning or localization problem, not a spend more problem. Those are different fixes. Speaker 4: Or you picked the wrong segment. Rachel: Exactly. And three, this is non-negotiable, fully loaded logistics cost as percent of revenue. Duties, returns, failure rates, all of it. Max: If you're hemorrhaging 20 to 25% of margin to the plane and customs, this isn't a corridor, it's a subsidy program. Rachel: Let me add one AI twist to each. On CAC, use AI agents to test creative and placement hyper-fast, but actually in language and in culture. Don't just ship your U.S. ads and hope. Max: Yes. Rachel: On repeat, plug your product and review data into models that tell you which SKUs resonate per region. Maybe Thailand loves bundle A, India loves refill packs. You'd be surprised what the data shows. Max: That's not just margin protection, that's the operator discipline that keeps units sane. A three-day promise at 95% confidence with shipping at 15% of order value is where you... After you actually make money on repeat. Rachel: And on logistics cost, use AI route selection and promise engines. Only show three day delivery where you can actually hit three day at ninety five percent confidence and keep shipping under fifteen percent of order value. Max: Right. AI is your operational optimizer for execution. But the decision to enter a corridor—China, EU, India, Southeast Asia—that's still a strategic bet. And if the numbers don't hit in six to nine months, you kill it and move on. Rachel: Yeah, here's the hard guardrail. If after six to nine months, your CAC is still double, repeat is lagging by 30%. And logistics cost is eating north of 25% of revenue? Max: You pause. Rachel: You pause. Don't fall in love with the story of a market that doesn't like your unit economics. Max: Or the product doesn't actually fit their market. Rachel: Also that. Max: Here's the hard truth: The next billion in GMV is absolutely there. India, Southeast Asia, China, EU, but only for brands with a solid ops foundation. First, not will fix ops later. Treat AI as a multiplier, not a magic fix. Use hard metrics, use a discipline ladder, and when the numbers say pause, you pause. Rachel: And if your house isn't in Order yet, that's fine. Fix your data, clean up your ops, get your payments Right. Max: Then go shopping for your favorite new corridor. Rachel: And that is the Checkout Point for this week. Alright, we're going to land this plane. If you remember one thing, it's that olive oil moment, hands covered in flour and AI just buying for Max: Mm Rachel: you. Max: -hmm. Rachel: That's where this is all heading. Max: Yeah, and your takeaway as an Operator is simple. AI that doesn't move a Real KPI, conversion, AOV, or margin is expensive theater. I've watched that movie at Amazon and Walmart too many times. Rachel: Real talk-if the bot doesn't actually convert better than no bot, you're just LARPing with someone else's margin. Max: Please put that on a dashboard-hard numbers, no spin-and map your next ninety days where AI actually helps you sell more, spend less, or reduce operational chaos. Not hypothetical-real. Rachel: Right. So whether it's Shopify's stack, your checkout mix, or that first market test, pick one bet, measure it hard. Numbers don't lie. Max: If this helped you stress test your operations or challenge an AI pitch, hit follow, subscribe, and drop a review. It helps other operators who care about margins find us. Rachel: Yeah, yeah. And seriously, send this episode to that one person on your team who keeps saying we need AI without actually building a spreadsheet. They need to hear this. Max: Thanks for hanging out with us today and asking the tough questions that matter. Rachel: This has been the Checkout Point. Max: See you next time! Rachel: Bye, everyone.