@crowizard_stef @blvckledge @vishalgoswami0 @Goldikam @dimiz1111 @marcdomancie @C_Shelefontiuk @mechellecarpio @maxlayn @f3dericobartoli 📷 𝗜𝗠𝗣𝗢𝗥𝗧𝗔𝗡𝗧: 𝗖𝗮𝗹𝗹𝗶𝗻𝗴 𝗼𝘂𝘁... 𝗮𝗹𝗹 𝗦𝗵𝗼𝗽𝗶𝗳𝘆 𝗣𝗮𝗿𝘁𝗻𝗲𝗿𝘀, 𝗦𝗵𝗼𝗽𝗶𝗳𝘆 𝗔𝗴𝗲𝗻𝗰𝗶𝗲𝘀 𝗮𝗻𝗱 𝗚𝗿𝗼𝘄𝗶𝗻𝗴 𝗗𝗧𝗖 𝗠𝗲𝗿𝗰𝗵𝗮𝗻𝘁𝘀 CRO (Conversion Rate Optimization) is the secret weapon for Scaling Shopify Brands in 2025! You’re a Shopify brand owner, a dropshipper, a marketing manager... You’ve scaled your store, and spent tens of thousands - maybe even millions on ads... Yet despite all that effort, your ROAS is shrinking. Acquisition costs are rising. And your store just isn’t converting the way it should. 𝗪𝗵𝘆? Because in 2025, traffic alone isn’t enough. More visitors don’t mean more customers. More ads don’t mean more revenue. The brands that are truly scaling this year are the ones mastering CRO (Conversion Rate Optimization) & Retention. 𝗙𝗜𝗡𝗔𝗟𝗟𝗬... After working with countless Shopify brands, we've learned that true growth isn’t about spending more, it’s about converting better. That’s why we’ve gathered top CRO & Retention experts for an exclusive event, revealing what truly drives conversions, enhances shopping experiences, and keeps your brand ahead. And now, it’s time to officially unveil: 📷 𝗢𝗙𝗙𝗟𝗜𝗡𝗘 𝗘𝗩𝗘𝗡𝗧: 𝗦𝗵𝗼𝗽𝗫: 𝗙𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗘𝗰𝗼𝗺𝗺𝗲𝗿𝗰𝗲 𝗘𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲 – 𝗖𝗥𝗢 𝗗𝗔𝗬 𝟮𝟬𝟮𝟱! 𝗗𝗮𝘁𝗲: April 17, 2025 𝗟𝗼𝗰𝗮𝘁𝗶𝗼𝗻: Hanoi, Vietnam 📷 𝗪𝗵𝗮𝘁’𝘀 𝗶𝗻 𝘁𝗵𝗲 𝗲𝘃𝗲𝗻𝘁? - Top industry experts breaking down conversion-boosting & funnel optimization strategies. - Successful case studies from brands that mastered CRO. - High-impact networking with Shopify brands, merchants, agencies & eCommerce builders. 📷 Stay tuned, follow us, and get ready for exclusive updates on speakers, agenda, and how to secure your spot! #ShopX #CRODAY2025

Coffee Chat #5: How about a buildathon? Last year in China, I kept getting pulled into small ecommerce coffee chats. A few people, a table, and hours of honest talk about building online. Then in India, I saw teams running buildathons. Everyone in one room, laptops open, building side by side. And I keep wondering: Why don't we have more of this for the ecommerce community in Southeast Asia, or more specifically, in Vietnam? So I started running a few here. We've run four 4S Ecom Coffee Chats so far, about vibe coding, building a mini app for your shop, sharper offers, UGC and email with AI. The first four were good conversations. But to be honest, sometimes talking and listening isn't enough - attendees want to walk out with something you actually built. Then, this time it's a BUILDATHON. The idea here is simple: Less watching, more doing. We take real problems from your own store and work on them right there, using AI tools that already exist. No theory, we spend the day building, not just listening. Here's the plan: Saturday, July 25, 09:00 a.m. to 05:00 p.m. GemCommerce office, 3 Le Trong Tan, Hanoi Lunch's on us, so you can stay and talk through the whole day What do you think about the format of a buildathon? If you're interested, message me or register here to save a spot: forms.gle/wZDyEyWq6iCXCD6QA You can also catch up on the other coffee chats, and sign up for this fifth one, here: chriskoh.dev/

Three coffee chats in, here's what we've learned Last year, I spent a few months hopping between coffee chats in Beijing and Shanghai. The experience there made me wonder: why hasn't this kind of event caught on more widely? So we started the 4S Saigon Ecom Coffee Chat. Well, three sessions in, here's a quick recap of what we've been up to. ☕ Coffee Chat #1: Vibe Coding with Claude Code for Shopify -We built three Shopify tools live: a CRO audit tool, a competitor auditor, and a Shopify theme block. -Rather than starting from a blank screen, we kicked things off with working versions and improved them together based on what people wanted to see. -One thing became obvious pretty quickly: you don't need to watch someone code for an hour. Starting with something that already works makes the whole session easier to follow, especially if you're not a developer. ☕ Coffee Chat #2: Build an App for Your Shop - This time around, we flipped the format completely. Smaller group, more hands-on, and everyone left building a mini app for their own store. - Projects ranged from review inboxes and content calendars to product description generators and simple CRMs. - Probably the biggest mindset shift? Claude works best as a junior developer. Your job isn't to argue with it, it's to review what it builds, explain what's broken, and keep moving. ☕ Coffee Chat #3: Sharper Offers with AI - Then we switched gears. No coding this time, we focused on offers instead. - After looking through 1,000 Shopify stores running ads, one pattern kept coming up. Almost everyone has an upsell widget, but very few have an offer that's actually designed for their own customers. - We were also joined by Artjom from MintFunnels and Pete Devkota from Email Optimize, who shared practical ideas around bundles, personalization, and matching the right offer to the right customer. If there was one takeaway, it's this: widgets aren't the advantage anymore. The thinking behind the offer is. AI just helps you test better ideas a whole lot faster. Looking back, every session has ended the same way. At the end of the day, it provides attendees something to bring home: a working tool, a live app or an offer to test. If you missed any of the sessions, you can catch up here: - Session 1 - Vibe Coding with Claude Code for Shopify: chriskoh.dev/gemlabs-03/ - Session 2 - Build an App for Your Shop: chriskoh.dev/gemlabs-02/ - Session 3 - Sharper Offers with AI: chriskoh.dev/gemlabs-03/ Anyway, that's three sessions down and three completely different problems we've explored together. Looking forward to seeing what we build at the next 4S Saigon Ecom Coffee Chat.

Been heads down researching 𝗖𝗵𝗶𝗻𝗮'𝘀 𝗲𝗰𝗼𝗺𝗺𝗲𝗿𝗰𝗲 𝗺𝗮𝗿𝗸𝗲𝘁 for the past 12 months. Shanghai, Beijing, Hangzhou, Guangzhou, Shenzhen. Actually being there, not reading reports. Took notes. Sharing them now. What I found: - The serious operators don't post publicly. They run closed WeChat groups with weekly coffee chats. The real knowledge is shared in rooms most people never get invited into. - The barriers to going global - content, language, store operations - are collapsing faster than anyone outside China realizes. AI is doing the work that used to require teams. - Shopify Plus is expanding into Hong Kong for a reason. The cross-border wave is earlier than the headlines suggest. One thing that surprised me: the ecommerce ecosystem in APAC is far more developed than the English-language internet gives it credit for. Japan alone is a top-10 Shopify market globally. Most people I know in Southeast Asia have no idea. Going to start sharing more of what I saw - the markets, the people, the patterns that don't make it into any report. If any of this is relevant to what you're building, stick around.

Should you still build a Shopify App in 2026? What 21,000 catalog slugs say 👇 I crawled the entire Shopify App Store with a tool I vibe-coded. Here's what the data actually says 🧵 - Only ~768 apps are verifiably alive - ~10,600 apps have ZERO reviews (zombie listings) - Top 500 apps capture 99.7% of all reviews - 86% of priced apps cost under $30/mo - yet only 8 apps charge $100+ It means you're really competing with ~1,600 serious apps and ~150 real incumbents. So - is there still room? Yes. Here are some wide open lanes in 2026: → Shopify Plus apps priced $99+/mo (47k+ merchants, barely any premium tools) → Orders & Shipping (lowest ratings = most unhappy customers) → Agentic post-purchase support (AI that works end-to-end, not just a chatbot label): Highest-rated category, smallest competitive pool, and the area where AI agents make the most economic sense. → Non-English market apps (Japanese, German, French merchants are underserved): Localized apps for non-English markets. Fewer than a dozen apps tagged with non-US/UK origin that also serve non-English merchants well. What I would NOT start in 2026: a new review app, a new generic popup app, a new generic dropshipping app, a new "AI product description writer". Key takeaway? The App Store isn't over. It's just much smaller and sharper than the headline number suggests! Full breakdown → chriskoh.dev/blog/inside-the…

Last week, Manus AI, v0, and Replit all launched Shopify integrations that let users spin up production-ready stores through conversational prompts. Wow, that's truly impressive! A few years ago, launching an e-commerce store meant hiring developers, buying themes, configuring apps, and spending weeks getting everything live. Now you can describe what you want and have a working store in minutes! But there's one problem: Most AI-generated storefronts look polished, but the conversion part remains a question. And I think the bottleneck in ecommerce has shifted from "Can I build this store?" To: "Can this page convince a complete stranger to trust me enough to buy?" The AI era makes CRO more important. When everyone can launch a store in 10 minutes, the advantage moves to whoever understands customer psychology, buying behavior, and conversion best :)

Claude building insane "landing pages" in 10 minutes is probably one of the most impressive AI demos I've seen recently (I tried myself, too)! Honestly, the output looked awesome: clean layout, stunning visuals, seamless motions, reasonable copy. The problem is that a page can look convincing and still convert terribly at the same time. That's because initially, AI is optimizing for "what a good product page looks like", while CRO is about understanding why a specific visitor still hasn't bought. When reviewing AI-generated PDPs, I keep seeing the same gaps: - Proof appears, but not in the right order. - Claims show up before trust is established. - Buy boxes miss critical microcopy. - Objections get addressed, but too late. - CTAs exist, but without enough context. None of these are design problems. They're decision-making problems. To fix them, you need to understand who's landing on the page and what's stopping them from buying. Are they skeptical? Price-sensitive? Comparing alternatives? Trying to understand if the product actually works? AI can't reliably answer those questions from a product URL alone. That's why I don't think AI is replacing CRO anytime soon. It still needs someone in the room who understands customer psychology, buying friction, and conversion strategy. The winners won't be the people using AI the most, but those combining AI speed with real conversion thinking!

Skills I've been practicing recently as an AI-first ecommerce operator Many people thought becoming more AI-natives meant learning new tools. Looking back, most of the skills I've been developing aren't really about AI. They're about thinking. 1. Problem Definition The biggest mistake I see people make with AI is jumping straight into solutions. "Build me an app." "Automate this workflow." "Generate content." The better question is: - What problem are we actually solving? - Who has this problem? - How often does it happen? - What does success look like? I've found that once the problem becomes clear, AI is surprisingly good at helping with the implementation. 2. Systems Thinking AI has made me think less about tasks and more about workflows. For example: Content isn't just content. To have good content, we need tons of steps before that: researching > ideation > drafting > feedback, etc. When you start seeing the entire system instead of individual tasks, opportunities for automation become much easier to spot. That's also how many of our internal tools started, by simply asking: Where does the workflow break & cost us time? 3. Product Thinking I started my career as a builder, so AI has actually brought me back to something familiar: thinking like a product manager. Not: "What feature should I add?" But: "What job is this tool helping someone accomplish?" All of the useful internal tools we've built started with a workflow problem, not a technology idea. 🗝️One tip for you: Don't automate everything. Find the bottleneck that's limiting your team's output or ROI, then start there. The highest-leverage workflow is rarely the most obvious one! 4. Communication The more I work with AI, the more I realize clear communication is becoming a competitive advantage. Working with AI feels a bit like managing a very capable junior teammate. If you're too vague, the output drifts. If you're too controlling, you limit what the system can discover. The sweet spot is knowing when to provide constraints and when to leave room for exploration. I've found that better outcomes often come from describing context, goals, and expectations clearly - rather than telling AI exactly how to do something. --- You know what? The funny thing is that none of these skills is new. AI just makes their importance impossible to ignore :)

The biggest mistake I see people make when starting with vibe coding: they pick a software idea instead of picking up a problem. After spending the last few months building internal tools myself and talking with ecommerce operators at our Coffee Chats, I've noticed a pattern. People often ask: "What should I build first?" The answer is usually not a SaaS, but something much smaller, like a simple website, a content collector, a tool that crawls a few pages and organizes information, a personal dashboard, a to-do app, a workflow that saves you 10 minutes every day, etc. Why? Because they're simple enough to teach you how software works. Most beginners think the hard part is getting AI to generate code. From my experience, the hard part is actually being specific: - Who is this for? - What problem are they trying to solve? - What information needs to go in? - What should come out? These questions sound simple, but they're actually the foundation of product development. In software, people often call these "user stories", descriptions of real situations that users experience. The better you can describe those situations, the better AI performs. That's why when training our teams or sharing with others, I've become less interested in prompts and more interested in problem definition. In order to do that, we're required to understand the workflow, the user, the data and what "good enough" looks like. Once the problem becomes clear, AI is surprisingly good at suggesting the technology. That's also why our recent Coffee Chats have become less about AI tools themselves and more about how ecommerce operators think through real business problems, workflows, and opportunities for automation. Drop a mes if you also want to join our next ECom Coffee Chat in Saigon :)