A big project is paralyzing for a precise reason: "redesign the onboarding flow" isn't something you can do. It's not an action; it's a mountain. You can't start a mountain. The entire art of getting large things done is decomposition — turning the unactionable mass into a sequence of concrete moves small enough that starting is obvious and finishing is frequent.
If you can't start it, it isn't a task yet
The test for whether something is a real task is brutally simple: could you begin it in the next ten minutes without deciding anything else first? "Improve the docs" fails — improve which docs, how, starting where? "Rewrite the installation section of the setup guide" passes. Anything that still requires a decision before you can move isn't a task; it's a task-shaped fog. Keep splitting until each piece has a clear, immediate first action, because a to-do you have to plan before you can start will sit untouched.
Slice for finishable units
Break work into pieces small enough to complete in a sitting or two. This isn't just tidiness — it's motivation engineering. A task that takes three weeks gives you no sense of progress for three weeks, which is exactly how projects quietly stall. Tasks you finish in a day generate a steady drumbeat of completion, and that momentum is fuel. The goal is a stream of small, real wins, not a single distant one you can't feel yourself approaching.
Surface the dependencies before they surface you
As you decompose, some tasks turn out to be locked behind others — you can't write the copy until the design is approved, can't test until the environment's built. Finding these before you start is the difference between a smooth run and a Tuesday where you discover you've been blocked for days. Lay the tasks out where the sequence is visible — as cards on an action board like Axtio — so the chain is obvious and you're always working something that's actually unblocked instead of ramming into a wall you could have seen.
Track the pieces so the whole stays visible
Decomposition creates a new risk: a scatter of small tasks with no sense of the whole, where you're busy all day and can't tell if the project is actually advancing. The fix is a single tracked list — a checklist in a tool like MyTeamTask where the project is the parent and the tasks are checkable steps beneath it. Now progress is countable: eleven of eighteen done, these three blocked. The big scary project becomes a bar that fills, and "are we on track?" becomes a question with a number for an answer.
Re-decompose when a task fights back
Some tasks reveal themselves as too big only once you're inside them — you start "rewrite the install section" and find it's secretly five sections and a broken example. When a task resists finishing, that's not a discipline problem to push through; it's a signal to stop and split it further. Chronic inability to complete a task almost always means it was under-decomposed. The willingness to break a stubborn task down again, mid-project, is what keeps momentum alive — and momentum, in the end, is the whole game.
