Don’t Be a Goldfish

Did you know that goldfish grow to fit the size of their environment? A goldfish in a small bowl stays small. Put it in a pond, and it can grow to be a foot long. Same fish, different container.

Work does the same thing.

Parkinson's Law holds that work expands to fill the time available for its completion. Give a task two hours and it takes two hours. Give it a week and, somehow, it takes a week. We rarely finish early. We find more to do.

This isn't laziness. It's a feature of how we're wired. When there's time left on the clock, it feels wrong not to use it. So we refine. We add. We polish. We revisit decisions we already made. And often, by the end, the thing isn't better. It's just more worked-over.

The trap is subtle because the extra effort feels productive. You're still working. You're still improving things. But past a certain point, each additional hour returns less and less value, and the cost isn't just time. It's the other things that don't get done because you were still polishing the thing that was already good enough.

One of the most effective defenses against this is timeboxing: deciding upfront how much time something gets, and holding that boundary. But even more powerful is defining what Done looks like before you start. In software development, we call these acceptance criteria: the specific, agreed-upon conditions that tell you when something is finished. A well-planned sprint uses both: time-boxed research tasks (spikes) capped by the clock, and implementation tasks (stories) capped by Done. When a team commits to what they can complete in a sprint, they're not just planning work against a calendar. They're creating their own boundaries to prevent endless iteration.

The discipline isn't to work faster. It's to size the container deliberately, define Done before you begin, and when you arrive, evaluate consciously rather than just continuing because there's still time on the clock.

Don't be a goldfish.