How to review a week of work on a Sunday in five minutes
· 6 min read
Five minutes is enough for a weekly review if two things are true: the data was collected during the week without your help, and you ask only a handful of questions in a fixed order. The review is fast because you are reading, not reconstructing. Here is the setup that makes Sunday short, and the five questions, one per minute.
Why five minutes and not an hour
The long weekly review has a good reputation and a poor survival rate. An hour on Sunday is an hour you will find a reason to skip by week three, and a review you skip teaches you nothing. Five minutes can survive a busy weekend, a cold, a holiday. It is also, honestly, about as long as the useful part of an hour-long review lasts; the rest is tidying.
The trade is that five minutes can only answer questions whose data already exists. You cannot reconstruct Tuesday in five minutes. So the work moves to the setup: arrange for the week to record itself.
Before Sunday: collect without effort
Three sources, each already keeping itself.
- The calendar. Meetings, with times. Already there.
- Your own notes, if you keep a line a day. how-to-keep-a-work-log-automatically-on-a-mac.html covers a version that costs one sentence an evening.
- App time. This is the one most people do not have, and it is the one that makes the review concrete. Punchcard sits in the menu bar, notices which app is in front through the day, and prints a paper receipt at your closing time: one line per app with the time, a day total, a stamp. On Sunday it prints a week roll with the week laid out. Nothing to start or stop, no projects, no account, no macOS permissions.
The week roll is the document the five minutes is built around. If you do not use Punchcard, the same questions work with Screen Time’s weekly view (System Settings, Screen Time, App & Website Activity), with the caveat that Screen Time does not keep the weeks for you, so question four gets harder.
The five-minute script
Set a timer if it helps. One question per minute, in this order.
- The total. Read the week’s total hours. Say it out loud. Do not judge it yet; just register whether it is near what you would have guessed. Most people are off by hours, and noticing the gap is worth the whole minute. See why-your-estimate-of-your-own-workday-is-wrong-by-hours.html.
- The top three lines. Which three apps got the most time? Is the top one the app where your real work happens? If the top line is a messaging app or a browser and you are a developer or a writer, that is the week’s headline.
- The odd day. Find the day whose total is furthest from the others. Was it long? Was it nearly empty? You usually know why within a few seconds, and that reason is either a one-off or the start of a pattern.
- Against last week. Put last week’s roll next to this one. Is the total drifting up? Is the same app climbing? One week is weather; two weeks side by side is the beginning of climate. Keep the rolls (PNG export into a folder by week) so this minute is possible.
- One sentence for next week. Not a plan, a sentence. “Close the messaging app until eleven.” “Stop at six on Thursday.” Write it on the roll or on a note you will see Monday morning. Then stop.
That is the review. If you have time left, do not use it; the habit depends on the review being short. the-sunday-roll-a-weekly-review-you-will-not-skip.html is about keeping the habit going past the first month.
What the roll cannot tell you
A week roll shows where the time went by app. It does not know what you did in those apps, which project the hours belong to, or whether the long day was productive or just long. Punchcard records app names only (never window titles, document names or URLs), has no projects or tags, and does not score the week. It also does not see time away from the Mac: meetings in a room, calls on a phone, thinking on a walk.
So the roll answers the “where” and your calendar plus the daily sentence answer the “what.” Trying to make the roll do both leads to staring at it for the hour you were trying to avoid. Read the numbers, ask the five questions, and let the meaning come from your own memory of the week, which the numbers are there to prompt, not replace.
Keeping the rolls
Two reasons to keep them. Question four needs last week, and in three months the stack of rolls is a record of a quarter that no memory could produce.
Export each week roll as a PNG (Punchcard produces a 1080 by 1350 image) into a folder named by year, with the file named by the week’s Sunday. Or copy the plain text into a monthly note. Either takes ten seconds and fits inside the five minutes if you count it as part of minute four. The underlying data also sits in a single SQLite file at ~/Library/Application Support/Punchcard/punchcard.sqlite, and Export CSV gets the totals out as plain text if you ever want a spreadsheet view of many weeks.
The month roll, which prints at month end, is the same review at a larger scale and takes about as long. counting-the-days-what-a-month-of-receipts-teaches-you.html covers what to look for there.
Questions
What if I missed a day of tracking during the week?
Read the roll anyway. A missing day shows as a short or empty line, and the other six still answer the questions. If the Mac was asleep at closing time, Punchcard prints the missed receipt late under the heading SORRY WE MISSED YOU, so the day is usually there, just delivered later.
Should I review on Sunday or Friday?
Friday closes the week while it is fresh; Sunday sets up the next one. Punchcard prints the week roll on Sunday, so that is the natural moment, but the script works on either day as long as it is the same day every week.
Can I do this for a team?
Not with this setup. Punchcard has no team features; each person’s roll is their own. People can share their receipts in a channel if the team wants to, and how-to-share-a-summary-of-your-workday-with-a-remote-team.html covers that, but the five-minute review is built for one person looking at their own week.