Flame Spiral

Flame Spiral

Home
Notes
Start Here
Sample Tool
Archive
About

When a Project Deserves to Die

If you asked when it ended, there would be no date.

Flamer's avatar
Nolan Saye's avatar
Flamer and Nolan Saye
Nov 02, 2026
Article voiceover
0:00
-7:32
Audio playback is not supported on your browser. Please upgrade.
Cross-posted by Flame Spiral
"Also written by Nolan Saye, for Flame Spiral: how to end a project on purpose instead of letting it drift. It is the creative side of Saturday's question, deciding when to quit before you start."
- Nolan Saye

Imagine that two projects end on the same afternoon.

The first ends because its creative has run out of belief, but no decision is made. One week the creative is working on the project; the next, they aren’t. After enough weeks, the project is simply no longer being worked on. If you asked when it ended, there would be no date. If you asked why, the reason would be a feeling: tiredness, doubt, or the pull of something more exciting. The creative would likely call that feeling the reason, even though it was only the mood of the day the project happened to stop, rather than a considered judgment about the project.

The second ends because its creative wrote down, months earlier, what success would look like and how long it would be given to appear. That window closed and the evidence had still not appeared, so the creative spends an hour harvesting what the project produced, writes a page about why it is ending, and closes it on a date they can name. If you asked why, there would be an answer about the project itself, not a mood.

From the outside, these endings look identical: both projects are over, and both creatives have moved on. The first creative failed because the project ended by drift, without a decision. The second has deliberately closed the project. The difference is not in the outcome, which is the same. It is in whether anything was learned, and whether the next project begins with more judgment than the last one did.

If your projects mostly end by drift, that ending can feel like failure. This essay proposes another form: a project can be ended on purpose, on stated evidence, after harvesting whatever value it produced. The practice is to make deliberate closure a real option.

The test for a deliberate ending

Start with the test that separates a project that deserves to die from one that is merely difficult: a project deserves to die when the evidence that would have justified continuing it has had a fair chance to appear and has not appeared.

Name what would justify continuing

The evidence that would have justified continuing has to be named before the window is defined, not after. What response, result, or change in the world or in you would make continuing the obviously right call?

If you cannot name it, you are not running a project; you are running a hope, and hopes have no ending conditions, which is why they can run for decades.

Annie Duke calls conditions like this kill criteria in her book Quit: a state you can observe and a date by which it must appear. What follows adds what to do once they trigger: give the evidence a fair chance, then harvest the project and record why it ended.

Give the evidence a fair chance

A fair chance means the evidence had a real opportunity to appear. It does not mean you felt done, were tired, found something more attractive, or discovered the project was harder than expected. A fair chance means giving the project a real opportunity for the evidence to appear, and enough time for its absence to mean something.

Know when absence counts

The final condition is that the window closes without the evidence you named appearing. This means that the condition you named must remain unmet within the time you set, after a fair chance. That is the state in which ending the project is closure rather than abandonment.

Set the window before the result

Whenever possible, set the fair-chance window before the result is known. A condition chosen afterward may still reflect legitimate context, but it cannot protect against shifting the standard to fit the outcome. Therefore, write down the original condition before the result, so any later revision is explicit rather than silently retroactive.

For a project you are unsure about, ask what evidence would justify continuing it with a clear conscience. Write that down, then set the date by which the evidence must have appeared. When that date comes, if the evidence is there, the project continues on something stronger than hope. If it is not, you will know what to do, and ending it will not feel like failure.

Recover what the project produced

A project that is ending may still have produced things worth keeping, even when its intended outcome is no longer worth pursuing. You may have gained a skill you did not have when you started. The project may also have produced a structure, such as an outline, method, or sequence, that could be lifted into another project. Keep the contacts too, the people you spoke to because of the project, and the materials, drafts, research, or design that was good even though the product it belonged to was not.

A failed project can also leave you with a better question, one you could not have asked before running it. Include that in the harvest if it is genuinely new and useful.

List everything in one sitting, before the project is closed, and put the list somewhere you will find it. A creative who skips the harvest loses out on the only value the dead project still holds. They lose the skill, the structure, the contacts, and the question. Without the harvest, the creator has to relearn in the next project what this one had already taught.

Record why it ended

Keep the ending record to one page. Include the date, the condition you set and did not meet, and the fair chance the project received, including what that involved. Record what you harvested and where it is. Add one line about what you would do differently if you began the project again today. Write that line as a note to a future self who may actually begin the project again, not as self-criticism.

A record turns the ending from an event that happened to you into a decision you have made. In two years, when a version of the same idea returns, you can reread the record and ask whether anything about the project has changed.

What remains after closure

What dies, then, is narrow: the project ends, its promise is released, and its slot is freed. Its attention rent stops too since the project no longer claims your attention.

What survives is everything the project left behind: your capability, the filed material, and the sharpened question. The person who ran the project is more capable of running the next one than they were when they began this one, precisely because it was closed on evidence rather than dropped on mood.

Wednesday for paid members: The Project Kill Test uses four columns, an evidence row, and an obligations line beneath the grid. It forces exactly one choice: continue, repair, park, or end. The fifth choice, letting the project drift on without a decision, is not on the grid.

Also from Nolan Saye, in The Judgment Relay: Decide When to Quit Before You Start, the same question asked of a business instead of a project.

Flame Spiral is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

© 2026 Liahtman LLC · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for writing, podcasting, and media founders.