The 20-Minute Weekly Review That Actually Keeps Me on Track

Most weekly review systems collapse for the same reason most diets do: they ask for too much, too often, and the first time you skip one you quietly stop forever. I abandoned about three of them before I figured that out.

The version I actually stick to survives busy weeks because it is deliberately small. It takes about twenty minutes, I run it Friday afternoon (Monday works too), and it needs nothing but a single document. Here is the whole thing.

Why twenty minutes is the point, not a limitation

A long review feels more responsible. It is not. The value is almost entirely in the habit of stepping back, and a habit that takes an hour gets skipped the moment a week gets heavy — which is exactly the week you needed it most.

So I capped mine. If it ever runs long, that is a signal something is wrong with the week, not the review — and that itself is useful information.

The four questions, in order

Order matters. I tried other sequences and this one flows best, because each answer sets up the next.

One: What actually got done? Not what was planned — what shipped. I write it down even when it feels small. Partly for morale (busy weeks usually produced more than they felt like), partly to calibrate. If "done" is consistently shorter than the plan, my planning is too optimistic, and that is fixable.

Two: What slipped, and is it dead or just delayed? This is the question I used to avoid, so I force myself to be blunt. For every unfinished thing I pick one word: delayed (still matters, moves to next week) or dead (quietly stopped mattering, let it go). The relief of formally killing a task I was never going to do is underrated. A list full of zombies is heavier than an honest short one.

Three: What is the one thing next week lives or dies on? Just one. Not priorities plural — the single outcome that makes next week a success even if nothing else happens. Naming it now means I protect it before the week fills with other people's urgencies.

Four: What annoyed me twice this week? This is my favorite, and the one nobody else seems to ask. Anything that irritated me two or more times — a clunky handoff, a file I kept hunting for, a meeting that should have been a message — is a system problem in disguise. One annoyance is noise. A repeated one is a process worth fixing. Most real improvements to how I work started as an answer to this question.

The part most people skip

Here is where reviews usually fail: people answer the questions, close the document, nothing changes, the review feels pointless, they stop.

My fix is one rule. Every review has to produce exactly one change for next week. Not five. One. Maybe I block two hours for the "lives or dies" task before anything else gets scheduled. Maybe I fix the annoyance from question four. One concrete change, written as a sentence I will actually see again.

One change a week sounds too modest. But fifty small, real changes a year compounds into a way of working that is noticeably better — without a single heroic productivity overhaul.

Doing it as a team without it becoming a meeting

If you run a small team, the trap is turning this into a status meeting, which is slower and less honest. The way around it: everyone answers the four questions in a shared doc before any conversation. Reading takes five minutes; the only things worth talking out loud are question three (so the team agrees on the one thing that matters) and any annoyance more than one person flagged.

If two people independently name the same friction, you have found your one change for the week — without a meeting that ate the afternoon.

When to change the questions

Treat the four questions as a default, not scripture. If a question stops earning its keep — you always answer it the same way, or it never leads anywhere — swap it. The structure is just a reliable way to make stepping back automatic. The specific questions should bend to your actual work.

What should not change is the size. The moment a weekly review feels like a chore is the moment it stops happening, and a small review you actually run beats a thorough one you abandon by February.

If you want a daily-scale companion to this, my guide on using ChatGPT for daily productivity covers the morning-and-evening loop that feeds into this weekly one.