I was not procrastinating. It felt like work, it looked like progress, and at the end of it the app was measurably nicer in a place almost nobody visited. Meanwhile the thing people actually got stuck on had been sitting in my notes untouched for a fortnight.
This happens constantly and it is not a discipline problem. There is a structural reason the wrong features feel more urgent, and once you see it you can catch yourself in the act.
Well specified beats important, every time
The fun features are fully specified in your head. You can picture the animation. You know which files to open. You know it will take a day and it will take a day.
The important ones are vague. "People do not understand what the report is for" is not a task. You cannot start it on a Tuesday morning with coffee. You have to think first, and thinking has no visible output, so it does not feel like working.
Given a clear task and a murky one, everybody picks the clear one. Not out of weakness. Clear tasks are how work normally gets done. The problem is that clarity and importance are unrelated, and in early products they are often inversely related, because the important things are exactly the ones you have not understood yet.
The other reason: the feature has no owner
When a real person asks for something, the request comes attached to them. You remember who wanted it and why, and if you skip it you are letting down someone specific.
Features you invented have no such anchor. They also have no one to disappoint if they turn out badly, which is precisely why they are more comfortable to work on. There is no risk of finding out you were wrong, because nobody is waiting to check.
How to catch it
One question, asked before starting rather than after: who asked for this, and what were they doing when the problem came up?
If you can name the person and describe the moment, build it. If you can only describe the feature, that is a signal. Not a prohibition, some things genuinely need to exist before anyone can ask for them, but a signal worth a minute of honesty.
The follow up that catches the rest: what would I have to see to decide this was a mistake? If nothing could tell you, you are not building a feature, you are expressing a preference, and preferences do not have to be shipped to be enjoyed.
The animation, in the end
I kept it. It is genuinely nice and it costs nothing to leave in.
But I have been honest with myself about the order. If I had spent that week on the thing people were confused by, I would have found out sooner that a section of the report was written for the wrong reader entirely, which turned out to be the biggest problem in the product and took a full rewrite to fix. Five days earlier would have mattered.
That is the actual cost, and it is never the feature you built. It is the fortnight the real problem sat there being vague while you did something well specified instead.
Appray walks you through the same territory in eight rounds of questions, then puts every feature you are considering on trial against what you actually said. The result is a Build, Kill or Prove verdict for each one, plus interview questions written from your own answers rather than a template.
It runs entirely on your iPhone or iPad. No account, no sign in, nothing leaves your device, and no AI is involved anywhere in it.