He Tried Another Developer and AI for His Shopify Store, Then Came Back to Me

With him since day one

He sells jewelry, mainly Tibetan accessories, and his annual revenue is in the tens of millions of yuan. In 2024, on the first day of his business, an existing client referred him to me to build a Shopify store. The store setup, the custom development, the apps, the theme: I did all of it.

So far it's been 1 project and 8 rounds of new requirements. The store grew from zero to tens of millions in annual revenue, and I've been with it the whole way.

The time I couldn't take it on

Once in 2026, he had new requirements. My schedule was full, and I couldn't take them on.

He didn't wait for me. He hired another independent developer.

Business doesn't wait. Not having time was my problem, not his.

Then the conversion rate collapsed

After that developer's changes, the store's conversion rate dropped from 28% to 2%.

Anyone in e-commerce knows what that means: the traffic keeps coming, the money doesn't, and every day is a real loss.

He came back to me. Once I took over:

• 10 minutes: restored the version of the source code I'd kept locally, so the store was up again
• 2 hours: went through it with him and disabled the plugins he didn't need
• 1 day: fixed the code problems and finished migrating the features

Speed came from rules

It wasn't that I'm especially good. My usual rules saved the day. When I work on a client's store, I stick to five rules:

• Changes to your store are made in a separate demo environment first. Only after you confirm everything is fine do I touch the live store. The live store is never a testing ground
• Every update is tested thoroughly before launch, and only goes out once it passes
• Every update is merged incrementally on top of the latest code. Nothing is overwritten, nothing is lost
• Every change is recorded and backed up
• All the code goes into a private repository. If you ever want to switch developers, every past version is there for you

You don't see these rules on a normal day. On the day something goes wrong, they're your insurance. A client's interests shouldn't rely on someone's character. They should rely on rules.

He tried AI too

He also tried building pages himself with AI. What came out didn't match the store's theme, and it didn't look good.

I'm not knocking AI. I use it for my work every day. This client is very good at marketing and has a great eye. He's someone who has really figured out selling. But running a business and building the tech are two different skills. AI can generate a page, but it doesn't know your theme and style, doesn't know which plugin will block payments, and doesn't know how to bring a store back in 10 minutes when something breaks.

AI removes the premium on writing code. It can't remove the premium on being responsible. He tested that for me, in a way.

What he says now

In his words: "Responds quickly, solves whatever comes up, good price, reliable."

Four phrases, and only one has to do with tech. When clients judge a developer, the focus isn't really on the tech.

The price is real, too: ¥600 a day, a public price sheet, unchanged for years. He tried someone else and he tried AI, so he knows what that price gets him.

After these years, I'm more and more sure of this: when business owners hire a developer, they aren't buying code. They're buying "someone handles it when things go wrong." Anyone can write code, and so can AI. Who answers your call when the store goes down at midnight is a different matter.

In the end

What makes this client typical is that he tried every option, another developer and AI, and then voted with his feet and came back.

No promise speaks as loudly as actual results.

If you remember one thing, make it this: keep the source code and documentation in your own hands. That's why he could be back up in 10 minutes. Anyone could pick it up quickly, because the key assets are in your hands.