Build journal25 May 20264 min read
Build journal — five weeks from beta
What's done, what's left, and the small decisions that consumed last week.
Five weeks from the beta open. Status from the desk.
What's done
- iOS app is feature-complete for v1.0. Trading mode, Recovery mode, Fitness mode all working end-to-end.
- The Apple Watch companion is streaming live HR and HRV cleanly. WatchConnectivity sync was the thing I was most nervous about; the last two weeks of dogfooding have been quiet.
- HealthKit pipe is in for Garmin, Whoop, Oura, Polar and Fitbit. Native Garmin is still a post-beta job — that £499/month API will only get licensed if the beta tells me there's enough Fenix/Epix demand to justify it.
- Beta tester guide, terms, agreement checkbox, and the public changelog page are all live on the site.
What's left
- Final pass on the contextual-baseline math. Want one more week of my own data through the model before it sees a tester.
- TestFlight build needs to clear Apple review on time. Submitting around 20 June to leave a real buffer.
- The first 100 invite slots will be filled from the waitlist by region + wearable mix — not first-come. I'd rather have 10 Garmin Fenix users than 100 Apple Watch users when the data question is "should we license Garmin native."
Last week's small decisions
- Pushed the beta date from 15 June to 1 July. I had a holiday I couldn't move; launching a beta while unreachable for a week is how a beta dies before it starts. Honest is better than rushed.
- Killed a section of the marketing site with fabricated testimonials. Even labelled as "stylised", invented quotes from invented traders is the wrong way to start. The site says "no testimonials yet, by design" now.
- Added name + country to the signup form. Now every confirmation email opens with "Hi [name]," which it didn't before, and I can see geographic mix in Firestore.
What I'm sitting with
Whether to do any paid promotion before launch, or whether to keep it word-of-mouth + Reddit + the studio's own audience. Leaning quiet. The first 100 testers matter more than the first 100,000 visitors.