Classify every captured item as an Action Step, Reference, or Backburner

When anything lands in your capture system, immediately sort it into one of three buckets — every item must belong to exactly one.

Why it works

Undifferentiated lists mix completable actions with reference material and vague future possibilities in a single visual field. This creates noise that makes it harder to identify what requires action, increases cognitive load during review, and allows important action steps to hide among inert content. Classification at the point of capture reduces the ongoing processing cost of review and ensures that action-requiring items are always identifiable.

How to do it

1. Action Step: any item that requires someone to do something — begin with a verb ("Send," "Write," "Decide"). Every Action Step should have a specific owner.
2. Reference: information you may need later — file it in a reference system, not your action list.
3. Backburner: ideas or tasks you want to revisit but not now — park them somewhere periodic review will surface them.
4. Classify before you close any meeting, conversation, or brainstorm — not later.

Evidence

Undifferentiated capture is a recognized failure mode in productivity systems: GTD’s "clarify" step addresses exactly this problem. The cognitive load basis for classification is consistent with cognitive psychology research on the cost of mixed-category lists. The three-bucket format itself is practitioner design. (mechanistic)

The underlying principle (distinguish actions from non-actions) is shared across multiple productivity frameworks; the three-bucket format specifically is not separately tested against others.

Common mistake

Classifying an item as a Reference when it is actually an Action Step deferred under a different label — the most common form of this is "I’ll revisit this" items that should read "Decide whether to pursue X by [date]."