Add To Cart is brought to you by
Were you forwarded this email? Sign up here to get the next edition.
I've spent a fair bit of time lately wondering whether the Shopify App Store is going the way of Limewire. If AI can connect directly to your store and just rewrite everything, does a third-party app still have a job to do?
Martin Cox is the go-to person for many when a technical Shopify question comes up. I’ve seen the Martin-effect in full force. He spent more than a decade on the agency and consultancy side before walking away at the end of last year to start building apps with the Native App Co. Four of his apps, all with animal names, are all live in the Shopify App Store. I asked him if it was a stupid time to build a new business building Shopify apps. Luckily, he didn’t walk out.
Cheers
Bushy
P.S. Martin is offering ATC listeners 3 months free on his app, Product Pelican. Use the code ADDTOCART at productpelican.nativeappco.com/addtocart
EPISODE CHEAT SHEET
⏭️ SKIP TO THE GOOD BITS
Why Martin quit agency life to build Shopify apps (~6:30)
Is it actually a stupid time to start an app business? The guardrails argument (~11:30)
Stork Credit: why store credit beats a discount code every time (~18:30)
Omnibus Owl, price history compliance, and what AU can learn from Europe (~24:00)
Fish Wishlist: what wishlist data actually unlocks for merchants (~26:30)
Product Pelican: your product data is your most important asset right now (~31:30)
The global catalogue, plain English, and the yellow t-shirt rule (~38:00)
Easy in, easy out: Martin's framework for evaluating any Shopify app (~47:00)
💬 QUOTE OF THE DAY
"Good data is the most important thing to make sure that your products surface well."
🧠 THE BIG LESSON
Apps aren't dying, they're becoming the guardrails
The gut instinct on AI is that it kills the app. If you can point Claude at your store and tell it to rewrite every product description, who needs a tool? Martin's answer reframes the argument. You can do that but it's cowboy behaviour with real risk attached. The value isn't the raw capability or transformation of data, it's the safe, repeatable interface around it. Here's how he thinks about building and buying apps in this new world.
Rule 1: Leverage what's out of the box first
Martin learned the hard way in agency days that trying to build enterprise everything on Shopify burns time and money. Start with native functionality, then only build custom where you genuinely need to.
"The best thing about SaaS is that you've got so much stuff out of the box."
Rule 2: Wrap AI in guardrails, don't hand it the keys
His app Product Pelican runs a product-data health check, then opens the AI prompts so you can set your own rules, like always mentioning free shipping, without letting an agent loose on your whole catalogue.
"Apps are really about creating the guardrails for people to do what's possible with APIs and AI."
Rule 3: Buy for easy in, easy out
The best apps drop in and pull out cleanly, connect to Shopify Flow, and come with people you can actually talk to. If onboarding takes two months or digs deep into your codebase, that's a red flag, not a feature.
"If your product's a good product, it doesn't need two months to onboard and set it up."
There are real commercials under all of this. Shopify has reported roughly an 8x increase in AI traffic to storefronts and a 13x increase in orders from AI searches year-on-year. Shoppers arriving from AI are also converting at nearly 50% higher rates and spending 14% more per order.
If AI is becoming a distribution channel, the merchants who win are the ones whose product data is clean, plain English, and structured. Capability and coding are cheap. The guardrails and the data will be the advantage.
✏️ ECOMMERCE ACTION TIPS
Replace one discount code campaign with store credit this month and track whether it pulls a second purchase. Your acquisition cost starts spreading across two baskets instead of one. (~22:00)
Run a product data audit before your next AI tool install. Check for missing Shopify taxonomy categories, thin descriptions, and blank image alt text. Product Pelican automates this. (~32:00)
Use Shopify Flow integration as your first quality filter when evaluating any new app. If the tool has its own Flow triggers and actions, it was built with the platform in mind, not bolted on top. (~47:00)
Email the support team of any app you are seriously considering before you install. The apps worth running are built by people who will get on a call with you before you commit. (~48:00)
Cull any app your team is not actively using and uninstall it properly. Unused tools in a Shopify store fragment attention and erode confidence in everything else in the stack. (~50:00)
💾 TECH MENTIONED IN THIS EPISODE
Stork Credit. Native App Co's Shopify app for automating store credit. Martin runs spend-and-save promotions, welcome credit offers, and B2B segment rewards through it without any custom development, including multi-currency via Shopify Markets.
Shopify Flow. Native Shopify automation. Martin uses it as both a quality signal when evaluating third-party apps (Flow integration means the tool was built with the platform in mind) and a way to build advanced store credit logic without additional tools.
Claude. Used for vibe coding prototypes and building quick one-off campaign sites, and cited as the tool that converted Martin from Gemini.
Grok. The AI model library currently powering Product Pelican's data generation, chosen for speed and a free API tier to start.
Matrixify. The bulk import/export tool Martin references from the days when wrangling Shopify product data was a headache.
🎧 OTHER EPS YOU MIGHT LIKE
Gavin Ballard from Submarine: What the Hell is Composable Commerce Anyway? | EP282
Deciding what to build custom versus buy off-the-shelf is the question every operator wrestles with. A Shopify ecosystem veteran breaks down composable, flexible and headless commerce without the fog.
Agentic Commerce Is Here: What Shopify's UCP Means for Ecommerce | EP597
If the global catalogue becomes the distribution channel, your product data is what gets you found. This one digs into UCP, agentic shopping and what to fix in the next 12 months.
Putting the AI in Product Details: The Okkular Story | EP196
The taxonomy gaps and missing attributes that Product Pelican surfaces are not new. Abhishek Vora built Okkular to address them using computer vision with auto-tagging from product images, building out synonyms, writing descriptions at scale. An older episode, but the problem it solves has only got bigger since.
💬 CONTINUE THE CONVERSATION
Plenty to argue about here. Is the Shopify App Store genuinely a smart place to build right now, or is Martin swimming against the tide? Jump in whether you're a regular or it's your first post.
Not a member yet? It's free to join. 600+ ecommerce operators talking shop every week.
