6 min read
I build iOS apps alone. Protecht Health is live on the App Store. WayPoint, a planner for students, is in beta and aims to launch on November 9. So this is not a victory lap. It is a list of what went wrong on the first one, and what I changed on the second.
What failed
Purchases quietly broke. When I reviewed the data in September, about 58 percent of the purchases people started in Protecht were failing. I had been spending effort on getting people to the app while many of the people who wanted to pay could not. The cause was invisible because the app did not record why a purchase failed.
I made too much content. The blog published daily, which AI made easy. Thirty daily posts brought in 58 search clicks, and about half of them came from a single page. Volume was never the fix.
Ads reached the wrong people. In one 30-day window I spent $282 on Apple search ads for 23 installs, a blended cost of $12.27 per install. Only one campaign had a sane cost per install, and it was the narrowest one. The broader campaigns mostly paid for curiosity.
People signed up and never reached the point of the app. The app’s job is to import health records. The admin dashboard showed 107 total accounts and 43 import events in the app’s entire history. Plenty of people created an account and stopped.
I trusted a number that was wrong. The homepage was counting every page load twice because the analytics tag was installed twice. I was reading traffic that was not real.
What I would do again
Ship small and watch it. Version 2.4 added event logging for the purchase failures. Seeing the reason for each failure turned a vague worry into a list of things to fix.
Read the numbers on a fixed day. I pull the metrics once a week and write down what changed. It is boring, and it is the reason the points above are caught at all.
Fix the leak before buying more traffic. I paused the plan to get more installs and worked on the places where people were already dropping out.
What I am doing differently with WayPoint
Three things, and I will report back on whether they worked.
- Beta testers before launch. WayPoint went into TestFlight beta first, with a feedback form and a separate form for problems, so I hear about breakage from a person before the App Store does.
- Decide what the app collects on day one. I wrote the privacy policy early and built the app to match it. No account, no server that stores schoolwork, and crash reports that contain no schoolwork.
- Write the failure down while it is fresh. This list exists because I kept notes from the first launch.
If you are about to ship something
- Log why things fail, not just that they failed.
- Do not read a traffic number until you have checked that it is counted once.
- Treat your first month of analytics as a bug report, not a scoreboard.
I do the same kind of work for small teams: finding where a process leaks and fixing the finishing, not adding more starts. If that sounds like your situation, the free one-hour session is the easiest way to see whether a Workflow Triage would help.
Where the numbers come from
- 30 daily posts and 58 search clicks: my own Search Console review, Sep 13, 2026.
- Agent briefs and the review step: my own workflow files for Protecht Health.
Work with me
Something keeps slipping?
A Workflow Triage names the three workflows worth fixing, in order, in one week. $1,500, credited toward a Pilot.