What Evidence Tells You a Learning Project Is Finished?

A project can keep expanding long after it has taught you something useful. But stopping too early can leave the lesson hard to see—or share.

For a learning project, what tells you it’s complete enough to close? Is it a working result, a clear record of what changed in your thinking, feedback from someone else, or something different? How do you leave room to improve the work without polishing it indefinitely?

I’m interested in the signal you use to close the loop and carry the learning into what you try next.