Most projects don't fail. They fade.

You probably have one. A side project you haven't touched in months, still marked "in progress" somewhere, maybe only in your head. You haven't worked on it, and you haven't let it go either.

Most projects don't end with a failure you can point to. They fade. And a faded project isn't free: it keeps a small slice of your attention, every time you remember it, feel a little guilty, and forget it again. David Allen has a name for this in Getting Things Done: open loops. Anything unfinished keeps pulling at you until you either finish it or close it on purpose.

I've let three projects fade this way. Here are the three ways it happened, the sign I missed each time, and how I close projects now.

It works, but nobody can use it

At sixteen, during COVID, I built a set of apps to replace my high school's paper attendance registers and timetables. The headmaster backed it from day one, and technically it worked.

Then it met the people who had to use it. Some teachers didn't own a smartphone. Some didn't understand why an app needed the internet, and a few told me it had eaten their phone credit, because they had never bought mobile data. When the paper registers came back, usage quietly slipped away, and I couldn't find anyone to hand it over to.

The sign I missed: I was explaining the product to people who had never asked for it. Building it was the easy part. Getting a teacher with a keypad phone to switch was the actual project. Adoption isn't the last step after building. It is the product.

You protect a name before you have customers

In 2024 I co-founded a small tutoring business teaching web development to university students. Two cohorts, and every student passed. Then people started asking for math and physics tutoring, and we had the network to serve them.

I said no. I had just learned about branding, I wanted a brand that could scale, so it had to stay strictly about web development. We had one cohort behind us and I was already defending a name. When the next cohort didn't appear on its own, neither of us went looking for it, and the business stopped moving.

The sign I missed: I was saying no to people who wanted to pay us. Follow the demand, not the name. A brand is what customers build for you after you serve them, not something you defend before you have any.

It only survives in bursts

In late 2024, Google's NotebookLM started turning documents into podcast-style conversations. I posted a story about how good the voices sounded, and the idea came right after: take a university course chapter, turn it into audio, split it into short parts, and lay it over Minecraft parkour or Subway Surfers footage. Brain rot, but for studying. I called it StudyRot. In about two months it covered six courses, and the best videos passed 16,000 views.

Gathering the material was easy. The editing wasn't: where each part starts and ends, captions, text over the footage. I worked in bursts, a heavy week and then nothing. Exams came, I paused "for now", and I never posted again. The project stayed marked as ongoing for more than a year.

The sign I missed: every restart cost more than the last. A week of work followed by three weeks of nothing doesn't average out; the project decays while you're away. If something only survives in bursts, it needs a smaller routine or a decision.

How to close a project on purpose

Stopping is a skill. Sometimes the right move is to keep pushing and find another way in. Sometimes it's to stop and start the next thing. The one thing not to do is leave it undecided. This is the checklist I use now; it's how I finally closed StudyRot this week.

1. Decide in one sentence. "I'm stopping this because …, and I'd restart it if …". If you can't finish the second half, it's closed, not paused.
2. Write a closing note. What worked, what didn't, the numbers, and exactly where you'd pick it up. Five minutes now saves a week of rediscovery later. It's also the only reason I could write this piece with real dates and real numbers.
3. Change its status where you'll see it. Archive it in your task app, so it stops showing up as something you owe.
4. Tell the people involved. A co-founder, early users, anyone still waiting on it. A clear "we're stopping" is kinder than silence.
5. Carry forward what transfers. StudyRot's core idea, helping students study the courses they actually take, became StudiPac, the study platform I run now. A closed project can still be a starting point.

Starting from the other end

These days I try to start where all three went wrong: with the people. Talk to the ones who have the problem, ask them to pay before building anything, and ship something rough early so they can tell you what's wrong. It doesn't stop projects from ending. It just means they end with a decision.

If you're still carrying one, close it this week. It takes an evening.