Writing

Build, Fix or Ignore: How to Decide Each Week When You Run Several Apps Alone

By Scott Wilder · October 20, 2026 · 7 min read · All writing

Once a week, put every app you run into one of three buckets: build, fix or ignore. Take at most one build and two fixes, and everything else waits. By the end of this post you'll have a 15 minute weekly routine, a scoring formula for the build slot, and a one-line template you can paste into your task list.

This is a suggested method to try. The scoring formula is a published framework. The weekly limits are my rule of thumb, so adjust them after a month of using them.

Fix anything that is broken before you build anything new

A broken app always wins. Fix means one of these is true right now:

Nothing else gets time until that is working. When two apps both qualify, fix the one with more users first. If you don't know the user counts, fix the one that touches payments or logins first.

Skip cosmetic problems. A misaligned button is not a fix item. If you fix small things while a real problem stays open, the real problem stays open longer.

Give the build slot to the one app users are pulling on

If nothing is broken, you build. Build means new work on a single app, the one where a real person asked for something or where people already use it most.

Not your favorite app. Not the one that's fun to open. The one with the pull.

Building on a favorite is the most common way to lose a week: you spend hours on work nobody asked for, and the app that had users gets nothing.

Ignore everything else on purpose, and write it down

Ignore is the bucket most people skip, so every app ends up with a little attention and none moves forward. Spreading yourself across every project is how the one that could have grown stalls, a point made in this Indie Hackers post on the hidden danger of side projects, which suggests pausing side projects to concentrate on a priority one.

The rule is simple:

Use a one-line template and a 15 minute timer

Do this at the same time every week. Don't open any code during the review.

  1. List every app with one line: name, what it does, the last time a user did something.
  2. Set a 15 minute timer.
  3. For each app, write one word: build, fix or ignore.
  4. Count. At most one build and two fixes. Move any extras to ignore.
  5. Paste the decisions at the top of your task list. Anything not on the list is parked.
  6. Next week, repeat.

Copy this line for each app:

[App name] | [signal this week] | [build / fix / ignore] | [reason in 5 words]

A signal is something that actually happened: a user message, a failed payment, an error alert, a signup. Zero signals and nothing broken means ignore. Any error or failed payment means fix. The app with the most positive signals is your first build candidate.

Here's what a filled-in week could look like. Example: say you run four small apps.

App Signal this week Bucket Reason
App A 1 failed payment Fix Money is leaking
App B 3 users asked for export Build Clear user pull
App C None Ignore No users, no requests
App D None Ignore Nothing happened

One fix, one build, two ignored. The whole list took five minutes.

Add a maintenance bucket for apps that are fine

An app that is stable, earning or growing a little, and needs nothing new can sit in maintenance mode. One Indie Hackers write-up on running several products at once describes a feature-complete maintenance mode with a cap of one new development a week.

Adapted for the weekly review, add a column called "mode" with three allowed values:

Only an active app can take the build slot. That one column stops a finished app from quietly eating your Saturday.

Pick the build with a four number score, not a gut feeling

When two or more apps compete for the build slot, score them. Use the RICE formula: Reach times Impact times Confidence, divided by Effort. Bubble's guide to RICE covers the four factors, and Project Wizards gives the same formula. A higher score means a higher priority. RICE was built for product teams, so here is a solo version that stays small.

Factor How a solo builder fills it in
Reach Users who asked for it or use that part of the app
Impact 1, 2 or 3, where 1 is barely noticeable and 3 is big
Confidence 50% if you're guessing, 80% if a few users asked, 100% if many did
Effort Work sessions, where one session is 3 to 4 hours

Use work sessions for effort because that's how a day actually goes. The same Indie Hackers write-up says its founders split tasks so each fits one working session of about 3 to 4 hours, one task on one product, with the rest of the day left for marketing, support and community.

Count the whole job, not only the coding. Testing, deploying and support take sessions too.

Example: say you have three build candidates.

Idea Reach Impact Confidence Effort Score
Export button for App B 10 2 80% 2 (10 x 2 x 0.8) / 2 = 8
New dashboard for App A 3 3 50% 1 (3 x 3 x 0.5) / 1 = 4.5
Dark mode for App D 4 2 80% 2 (4 x 2 x 0.8) / 2 = 3.2

The export button gets the slot. Write the score next to it so you can see later whether your guesses were right.

Four rules keep the score honest:

  1. Never score a fix. A broken app already won. Scoring it wastes the time you're trying to save.
  2. Keep the scales fixed. If you change them every week, scores from different weeks can't be compared.
  3. Keep confidence low until a real user asks. Otherwise the idea you like wins on a guess.
  4. Skip the cutoff. Rules like "above 100 is a go" come from product teams with different scales. Rank your own candidates against each other.

If the top two scores are close, take the one that fits in a single session. If the top item has confidence of 50% or less, make the build a small test first, such as asking users or putting up a simple page, before you write real code.

Size the build to fit one session or cut it down

Before anything goes in the build slot, ask: does this fit in one 3 to 4 hour session? If not, what is the smaller piece?

Cut it down before it goes on the list. A build that doesn't fit in a session turns into a fix next week, because half-finished work breaks things.

If you only have evenings, the schedule in How to Ship Side Projects While Working a Full-Time Job shows how to fit those sessions around a day job.

Shut an app down after four weeks of nothing

An app that sits in ignore for four weeks in a row, with no users and no requests, needs a decision. Either keep it parked or shut it down. Leave a note beside it so the review doesn't repeat the same question forever.

Never shutting anything down is its own cost: the list grows and the weekly review takes longer each time.

When you park or close an app, write one line:

[App] | [date] | [why] | [what would make me reopen it, e.g. 3 user requests]

The Indie Hackers advice on running three projects goes further: commit fully to one project for about 4 months, then decide to continue or shut it down, instead of giving it a small share of effort for years. If you can't tell whether an app deserves more work, that is a cleaner test than another year of ten percent attention.

Let the rule pick your week, not your app ideas

The rule doesn't decide what a good app is. It decides what you touch this week. If you want the bigger picture of how I run several apps with a team of agents, How I Run a Portfolio of Small Apps With a Team of AI Agents covers it.

Skip the complicated scoring sheet. You'll spend the weekly time keeping the sheet up to date instead of deciding. The sheet is four numbers and one line per app.

Your next ten minutes

  1. Open a note and list every app, one line each.
  2. Next to each one, write build, fix or ignore.
  3. Count: one build, two fixes at most. Move extras to ignore.
  4. Put the result at the top of your task list.

That's it. Do it again next week, same time.

Got a task that keeps landing in your fix bucket?

If one manual task keeps coming back every week and you'd rather have it automated, tell me what it is in one sentence and I'll tell you if it's worth building. Email me.

Questions

How do I decide which side project to work on this week?

Fix anything broken first. If nothing is broken, give the one build slot to the app where a real user asked for something, or where people already use it most. Everything else is ignored until next week.

Should I fix bugs or build new features first?

Fix first when an app is down, losing data, charging wrongly or blocking sign-up. Cosmetic problems are not fixes. When nothing is broken, build.

When should I shut down a small app?

If it has sat in ignore for four weeks in a row with no users and no requests, decide to keep it parked or shut it down. Write one line with the date, the reason and what would make you reopen it.

Is RICE too much work for a solo founder?

Not if you keep it to four numbers per idea and no more than five candidates. Use it only to pick the build slot, never for fixes, and keep the scales fixed from week to week.

What is maintenance mode for a small app?

It is a stable app that needs nothing new. It gets time only if it breaks, and at most one small change a week. Only apps marked active can take the build slot.