It's Friday, late afternoon. You were busy all week — you're sure of that much. Ask yourself what went on Tuesday and the honest answer is a shrug.
That shrug is where most end-of-week reviews begin. You clear the inboxes, tidy the lists, and close the laptop having learned nothing about the week you just lived. By March the habit is gone.
This is the weekly review I'd hand a friend: the GTD steps on a fixed time budget, with one block in front that looks at what actually happened and one block after that turns it into a lesson.
Short answer: a weekly review is a 30–45-minute appointment with yourself, at the same time each week, to see where the week went, get your lists and calendar current, and decide what next week is for. Run it in five timed blocks: the record of your week (10 min), Get Clear (10), Get Current (15), three written reflection questions (7), and choosing next week's priorities (3). Start from a record, not from memory — the structured debriefs that have been studied work from specific events and outside data, and that's the step most templates skip.
A weekly review does two jobs. Maintenance makes your lists true again, because they go stale within days: finished actions stay unchecked, emails turn into promises that live nowhere, meetings spawn follow-ups. Learning catches the patterns a week is just long enough to show — the meeting-heavy Tuesday, the estimate that was off by half. A review that only does maintenance repeats the same week forever. GTD covers the first job, the research on reflection and debriefs covers the second, and the template below runs both on a clock.
The GTD weekly review, as David Allen defines it
The GTD weekly review is an 11-step checklist in three phases — Get Clear, Get Current, Get Creative — and David Allen calls it the "critical success factor" of the whole method (GTD podcast #314, 2025). From the official weekly review checklist (David Allen Company, 2006):
- Get Clear gathers what's loose: collect stray papers, process every inbox to empty, and write down anything still circling in your head.
- Get Current walks every list: action lists, the past calendar (for leftover actions), the upcoming calendar, Waiting For, every project (each needs at least one next action), and any other checklists you keep.
- Get Creative looks outward: skim Someday/Maybe for anything that has become live, then add new ideas — the checklist explicitly invites "hare-brained" ones.
Its job is trust: once every open loop is written down with a next step, your head can let go of it.
There's lab evidence for that part, though not for GTD itself. In Masicampo and Baumeister (2011), unfinished tasks kept pulling attention away from an unrelated reading task until people wrote a specific plan for how, when and where they'd finish them. Listing several possible ways to finish, without committing to one, didn't quiet the thoughts — and planning stopped the intrusions without making anyone less anxious about the task. These were short experiments with students and online participants.
What the checklist doesn't have is a step that asks how the week actually went. It reads your calendar for loose actions, not for lessons.
Why reflecting on the week works — and what the evidence doesn't show
Reflection is the part most checklists leave out, and it has the most evidence behind it.
Reflection beat extra practice in a call center. Di Stefano, Gino, Pisano and Staats (2016) ran a field experiment with 101 trainee technical-support agents at Wipro in India. From the second week of a four-week training program, one group used the final 15 minutes of each training day to write down what they'd learned — the instruction asked them to "reflect on and write about at least two lessons. Please be as specific as possible." The control group kept working through those minutes.
On the final competency exam the reflection group scored about 21% higher on average, per the authors' 2016 presentation (HBS Working Knowledge (2014) reported 22.8% for an earlier analysis). Follow-up online experiments credited two things: a better grasp of the task and more confidence in doing it.
The same project found that people undervalue it. Offered the choice, 210 of 256 online participants picked more practice over reflection, a preference the authors say "does not seem to be well-informed."
Structured debriefs help, and they're short. Tannenbaum and Cerasoli (2013) pooled 46 samples covering 2,136 people. On average, debriefs improved performance by about 25% over controls (d = 0.67), and individuals benefited as much as teams. The debriefs averaged roughly 18 minutes, and session length had no significant relationship with how well they worked.
Their definition is the useful part. A debrief, for them, has four ingredients: you do the thinking yourself; the intent is to improve, not to judge; it's about specific events rather than general strengths and weaknesses; and it draws on more than your own impression — a teammate, an observer, or objective data. By that definition a reflective diary on its own doesn't count.
Hold the usual weekly review against those four ingredients and the weak one is obvious: it runs on impressions. Step 1 below supplies the outside data and the specific days, Step 4 is where you do the thinking, and the questions are worded to improve rather than to grade.
What this doesn't show. I couldn't find a controlled trial of the GTD weekly review itself. The reflection studies are mostly training or simulated settings, people learning something new.
The Di Stefano paper is still unpublished — Di Stefano's research page lists it as in preparation for resubmission — and its latest version reports that reflection pays off most early in the learning curve (UNC Kenan Institute, 2023). In the meta-analysis, debriefs without a facilitator, which is what a solo review is, showed a much smaller effect (d = 0.25) from too few studies to settle anything. Good reasons to reflect on specific events against real data, then — not a promise of a 25% better you.
Why weekly reviews die
Three ways a weekly review dies, and a fix for each.
Skipped when busy. Tiago Forte reports that in an informal poll of his own followers — a "productivity-obsessed audience", in his words — only 38% said they regularly completed a weekly review (Forte Labs, 2020). A well-supported fix is an if-then plan. Across 94 tests, forming one had a medium-to-large effect on reaching goals (d = 0.65) in Gollwitzer and Sheeran (2006). "When my last call ends on Friday, I open the review" beats "I'll do it sometime Friday."
Too long. Leo Babauta wrote that his GTD reviews often ran two to three hours (Zen Habits); Todoist recommends "a full hour at least" (Todoist). A review that eats an afternoon loses to the first deadline. Since debrief length didn't predict benefit, give each step a budget and move on when it's spent.
Run from memory. To say where the week went from memory, you'd have to recall every episode of an activity across seven days and add them up — a calculation Robinson et al. (2011) describe as hard even for people with regular hours, which is why time-use researchers lean on diaries.
They found people's estimates of their working hours ran 5–10% above their diaries; BLS economists Frazis and Stewart (2014) dispute the size of that gap. Those are survey averages about usual hours, and one person's memory of one particular week can miss low instead. When I read my own record every evening for a 30-day time audit, the biggest surprise was that I'd worked more than I remembered, and away time — calls, meetings — ran consistently longer than my memory of it. Put a record on the table before you judge the week.
A weekly review template: 45 minutes, starting from what actually happened
Here's the whole thing on one screen. The minute budgets are mine — a starting point to adjust, not a number from a study.
| Step | 45 min | 30 min | The question it answers | What to look at |
|---|---|---|---|---|
| 1. The record | 10 | 5 | Where did the week actually go? | Time by category; heaviest and most scattered day; this week vs last; unlogged away time |
| 2. Get Clear | 10 | 5 | What's still loose? | Every inbox to empty; a written mind sweep |
| 3. Get Current | 15 | 12 | Is every list true again? | Action lists; last week's calendar beside the record; the weeks ahead; Waiting For; projects |
| 4. Reflect | 7 | 5 | What did this week teach me? | Three written prompts, 1–3 sentences each |
| 5. Choose | 3 | 3 | What is next week for? | Someday/Maybe skim; three outcomes; first step of each on the calendar |
The 30-minute version works if you already clear your inboxes daily, which is Babauta's first tip. If all you need is a GTD weekly review template, Steps 2, 3 and 5 are it, with times attached.
Step 1 — The record (10 minutes)
Four things, in this order:
- Time by category. Does the split match what you meant this week to be about?
- The heaviest day and the most scattered day. One shows where your time went; the other shows where it leaked.
- This week against last week. A total on its own means little. Next to your own normal, it starts to mean something.
- The away time nobody logged. Calls, meetings, the hallway conversation. Name them now, while you still remember them.
I make Timex, and this is the step where it does the work. Pick Last 7 or This week and the week view draws one bar per day stacked by category, with a dashed line at your daily average. Under the total, a grey line reads something like "1h 20m more than the previous 7 days", compared over the same elapsed span. Hover a day for its total, Focus Score and top three categories; click it to open its 24-hour timeline, where a scattered day shows as a run of short segments in changing colors. The holes in that timeline are mostly away time — turn each one you can name into a manual entry with the + in the Today header. With the local MCP server switched on, you can also ask Claude or another connected client to summarize the week against last week.
Without Timex: on a Mac, System Settings → Screen Time → App & Website Activity can show last week by app or by category (Apple Support). Your calendar is the other free record — it shows the meetings, not the work between them. Either beats memory.
Step 2 — Get Clear (10 minutes)
Empty every inbox you have — email, saved chat messages, the notes app, Downloads, the paper tray — by deciding what each item is, not by doing it. Anything that needs real work becomes a next action. Then spend two minutes writing down whatever is on your mind that isn't captured anywhere yet.
Step 3 — Get Current (15 minutes)
Go through your action lists and tick off what's done. Read last week's calendar with Step 1's record beside it: the meetings that ran long and the follow-ups you promised are both visible now. Look at the next few weeks. Check Waiting For. Then walk the project list and make sure every project has a next action specific enough to start. Masicampo and Baumeister's finding is the test here: "work on the report" leaves the loop open; "draft section 2 on Tuesday at 9" closes it.
If Step 1 showed a task took twice what you planned, write down the ratio. That's raw material for beating the planning fallacy next time you estimate.
Step 4 — Reflect (7 minutes)
This is the weekly reflection block. Write the answers down, one to three sentences each.
- What are two specific lessons from this week? Borrowed from the Wipro prompt. "Meetings ran over" isn't a lesson; "the Thursday call needs an agenda or a hard stop" is.
- What went well, and why? In Ellis and Davidi (2005), soldiers on a series of navigation exercises improved more when each day's review covered successes as well as failures than when it covered failures alone. The causes of a good day are worth knowing too.
- What's one thing I'll do differently next week — and exactly when? Write it as an if-then: "When I open my laptop on Monday, I'll block the first two hours before checking chat."
Keep the tone the debrief research asks for: improve, don't grade. A heavy week might mean a demanding client, not a personal failure.
Step 5 — Choose (3 minutes)
Skim Someday/Maybe for anything that has become real. Pick three outcomes for next week and put the first step of each on the calendar. That's GTD's Get Creative, cut down to what fits in three minutes.
The honest limits
A record shows time, not value. Four hours in the editor can be your best work of the month or a long detour. The data tells you where to look; the reflection decides what it meant.
Timex sees your Mac and nothing else. Phone calls, in-person meetings and thinking away from the desk show up as gaps until you log them. It's Mac-only, with no phone app and no calendar integration, and it doesn't plan your week — Steps 2 to 5 are yours.
The evidence is suggestive, not specific. Reflection and debriefs have good support, mostly in training and simulated settings. This template is untested, and I couldn't find a trial of the GTD review against doing nothing at all.
Start next Friday
Pick the slot and write the if-then today: "When X happens, I open the review." Run the 45-minute version once, with a timer, and notice which step overruns.
If you'd like Step 1 to fill itself in, download Timex. The 100-hour trial is about two and a half work weeks — enough for two reviews, the second with a real week-over-week comparison.
— DK