Twenty years of building things. Most of them for other people. On Monday the 13th, Shopify approved Proviand. For the first time there’s something in a real app store that’s mine.
I’m proud of it. I also got there the stupid way.
Three tries
Submitted. Rejected. Fixed it. Rejected again, for exactly the same reason.
That second one stung, because it was my own fault. Shopify had changed the rules on how apps are allowed to log in. Mine hadn’t caught up. Everything that talked to Shopify broke. Everything that didn’t kept working perfectly. So it looked like two unrelated problems and I went chasing the wrong one.
Here’s the part I want to remember. I fixed a real bug and resubmitted without ever checking whether it was their bug. It wasn’t. I’d guessed, shipped and hoped. Which is exactly the thing I tell clients not to do.
Third time it went through.
Proviand looks at what a shop has already sold and tells the owner what to order next. That’s the whole idea. Shopify is switching off their own version of this on August 31, so there are a lot of shop owners about to need somewhere else to go.
Then came the sobering bit. Getting approved is not the same as getting found. I have a listing and almost no installs. Nobody is going to trip over it by accident. That part is on me and it starts now.
Koppel is waiting on Bol
Koppel connects a Shopify shop to Bol. Stock, orders, returns, all of it kept in sync so you’re not copying numbers between two screens.
The code has been finished since the end of June. What it was waiting on was Bol themselves, who wanted to check my company registration first. Nine days of nothing. Then last Thursday evening it landed: registered at one, seller account approved by eight. Saturday morning they sent a welcome mail that ends with “Tijd om te knallen.”
I’m still not submitting it yet. Proviand got rejected for a mistake Koppel might well share. I’d rather spend a day looking for it than another two weeks in the queue. Once bitten.
The green
I rebuilt the look of btng.studio on Friday. One long day.
The colour had bothered me for months. Every version came out too loud and I assumed I just had bad taste. Turns out the tool I was using to build the palette expects your brand colour to sit in the middle of the range. Mine is very dark, so it kept brightening it and then cranking it up to keep it looking green.
Months of blaming myself for something a slider was doing.
Fixed now. One green, one accent and a rule I should have written down years ago: the bright colour is for small things. Never for the button you actually want people to press.
The thing I didn’t want to write down
Friday I went through my sales pipeline. Twenty open deals became six. Fourteen of them had been quietly dead for months and I’d been rolling them forward, telling myself one more follow-up would do it.
Closing them was the work. Nudging them again was the story I’d been telling myself.
While I was in there I noticed something I’d never actually written down anywhere. My biggest client (the one currently paying for most of my life) came through a partner. So did the best deal still open.
Meanwhile: cold email, dozens sent, nothing live. LinkedIn, same. I’d even convinced myself outbound was working, because I’d been counting replies as if they were opportunities. They’re not. A reply is someone being polite.
So the one channel I never think about has quietly paid for everything. The two I worry about constantly have produced nothing.
I don’t know what to do with that yet. But writing it down felt like the point.
Two weeks. One app in the store, one waiting on a Dutch warehouse, one green fixed and one uncomfortable thing on paper.
Good fortnight.