How to Do a Weekly Review of Your Work (2026 Guide)

Block 45 minutes at the same time each week. Gather a record of what happened, sort it into a few buckets, compare it with your plan, write five lines and set three priorities. Toggl suits people who already run timers for client billing. Timing suits billable Mac work. ActivityWatch is free and cross-platform. Dayflow is free on a Mac and writes the timeline for you without timers.

A good weekly review leaves you with three things: an honest picture of where the week went, a short list of what actually moved, and next week’s priorities already on the calendar. Most reviews fail at the first one. You open a blank doc on Friday afternoon and try to rebuild five days from memory. Memory keeps meetings, deadlines and whatever annoyed you. It drops the quiet hours of real work, so the review becomes a list of what felt loud.

The fix is to collect the record first and judge it second. This guide walks through a six-step method that needs only a calendar and a text file. Software can take over the collecting step, so we also compare a few Mac tools that do it.

Disclosure: Dayflow is our product, and it is one of the tools covered. The comparison is based on vendor documentation: facts about the other tools come from each vendor’s own pages, checked on October 4, 2026, and Dayflow’s facts come from its 2.2.0 release.

The short version

1. Book the same 45-minute slot every week and protect it.
2. Gather a record of what actually happened, from sources other than memory.
3. Sort the week into four to six buckets and total each one.
4. Compare the totals and outcomes with what you planned.
5. Write a five-line summary: shipped, moved, stuck, learned, next.
6. Pick next week’s top three priorities and put them on the calendar.

What you need before you start

• Your calendar for the week under review and the week ahead.
• One running document for every review. Keep a single note with the newest week at the top rather than a new file each Friday. That way you can scroll back and spot patterns.
• Last week’s plan, if you made one. If you didn’t, this first review is where you start.
• Your trail of evidence: sent email, chat threads, commits or merged pull requests, closed tickets and files you edited.
• Optionally, software that records the week for you. The choices are covered in Step 2.

If you have never looked at your week as numbers, the where does my week go calculator gives you a rough starting estimate to check against the real record.

Step 1: When should you do a weekly review?

Pick a recurring 45-minute block and put it on your calendar as a meeting with yourself. Friday afternoon suits people who want to close out the week. Monday morning suits people who plan better with a fresh head, though by Monday the details of Wednesday have faded.

The exact time matters less than the repetition. Two months of short weekly reviews will tell you far more than one heroic three-hour review in January.

Illustration: a product designer books Fridays from 3:30 to 4:15, titles it “Weekly review” and marks it busy so nobody drops a call into it.

Mistake to avoid: treating the slot as optional. The first time it gets moved for a “quick sync,” it starts to drift, and within a month it is gone. If you must move it, move it to another time on the same day.

Step 2: How do you find out what you actually did this week?

Before you write a single opinion, collect the evidence:

• Calendar: list the meetings you actually attended.
• Sent mail and chat: skim for threads you drove.
• Task board and repo: note what closed and what merged.
• Untracked work: note anything that took an afternoon but left no trace, like a long debugging session or a document you read three times.

If you write a short note each day, this step gets much faster. The daily work log template is built for that. You write a few lines at the end of each day and read them back on Friday.

Software can do most of this collecting for you. Most Mac time trackers fall into two camps.

Timer apps record what you tell them. Toggl (formerly Toggl Track) is built mainly around timers and timesheets: you start a timer or fill in a timesheet. Its desktop app also offers automated background tracking and a timeline. See our Toggl for Mac comparison.

Automatic trackers record what the computer sees.

• Timing records apps, websites and documents on the Mac, down to specific URLs and Mail threads. See our Timing comparison.
• Rize tracks apps and websites, uses AI to categorize and tag them, and can pull meeting time from Google Calendar. See our Rize comparison.
• ActivityWatch records the active app, the window title and away-from-keyboard time. Browser extensions add tab titles and URLs. See ActivityWatch vs Dayflow.

Dayflow takes a different route. It records your screen and has AI write the day into a readable timeline. The record reads like “reviewed the onboarding copy, then debugged the export bug” rather than a list of app names. For a weekly review, export the date range as Markdown from Settings → Export and paste it into your review doc. The export work history workflow shows the steps.

Illustration: a developer’s calendar shows eight hours of meetings, and the commit log shows two merged pull requests. Their daily notes show that most of Tuesday and Wednesday went to a migration that never made it onto the task board. Without that third source, the migration would be missing from the review.

Mistake to avoid: rebuilding the week from memory alone. Whatever source you use, the record should exist before you start judging it. Our guide on how to remember what you worked on goes further on this.

Step 3: How should you categorize your work time?

Group the record into four to six buckets and total each one in hours, even rough ones. These buckets are a good start for most knowledge workers:

• Deep work on the main project
• Meetings and calls
• Communication (email, chat and reviewing other people’s work)
• Admin and small tasks
• Interruptions and unplanned work

Use the same buckets every week so the numbers stay comparable. Change a bucket only when it stops telling you anything.

Illustration: a consultant totals a 40-hour week as follows:

• 12 hours of client deliverables
• 11 hours of meetings
• 8 hours of communication
• 5 hours of admin
• 4 hours of unplanned fire drills

The meetings number is the one worth questioning. The meeting cost calculator turns a recurring meeting into a dollar figure, which makes the “do we need this” conversation easier.

Mistake to avoid: too many categories. Twenty project tags look precise, but they turn the review into accounting. If you bill by project, keep the detailed tags in your timesheet and use broad buckets here. If you want to dig into categories properly, run a separate time audit instead of bloating the weekly routine.

Step 4: How do you compare the week with your plan?

Put last week’s plan next to the totals and the list of outcomes. Then ask three questions:

1. Did the top priorities get the time they needed?
2. What took longer than expected, and by how much?
3. What showed up that wasn’t in the plan, and was it worth it?

Patterns show up in the second question over several weeks. Say you plan “write the proposal” as half a day and it takes two days, three weeks running. Then the estimate is the problem, not your effort. The time estimation test is a short exercise that compares your sense of time with a clock. It is useful if you suspect your estimates run off.

Illustration: the plan had “finish pricing page draft” as priority one. The record shows 3 hours on it and 9 hours on unplanned support escalations, and the draft didn’t ship. The honest note is not “I failed at the pricing page.” It is “support ate the week, so I need either a support rotation or a smaller plan.”

Mistake to avoid: grading yourself instead of the plan. A weekly review exists to adjust the system. If every review ends with “try harder,” you will stop doing reviews.

Step 5: What should a weekly review summary include?

Write five short lines in your running doc:

• Shipped: what is finished and out the door.
• Moved: what progressed but isn’t done.
• Stuck: what is blocked, and on whom.
• Learned: one thing you would do differently.
• Next: the headline for next week.

If you report to a manager or client, this summary becomes your status update with light editing. The weekly status report template has the structure ready.

Some people paste their exported Dayflow Markdown into an AI assistant to draft this. The Claude weekly report workflow walks through it. Two cautions apply:

• AI summaries can miss or misinterpret work, so check every line before you send it.
• If you give your data to a cloud assistant, that conversation is not local, even if the data was stored on your Mac.

Illustration:

• Shipped: export bug fix.
• Moved: billing migration, about two-thirds done.
• Stuck: pricing page, waiting on legal copy.
• Learned: keep support to mornings.
• Next: finish the migration.

Mistake to avoid: listing activity instead of outcomes. “Had 14 meetings” is activity. “Agreed launch date with sales” is an outcome. The summary should hold outcomes, since the bucket totals already hold the activity.

Step 6: How do you set next week’s priorities?

Choose at most three things that would make next week a success. Then open the calendar and book time for each one. Base the time on what Step 4 told you about how long things really take.

Illustration: priority one is “finish billing migration,” estimated at 8 hours. Last week’s migration work ran over, so the developer books 10 hours across Tuesday and Thursday mornings. They decline one optional meeting to make room.

Mistake to avoid: a list of ten priorities with no time attached. A priority without a calendar block is a wish. If the three blocks don’t fit, the plan is too big, and it is better to find that out on Friday than on Wednesday.

Doing it by hand vs with a tool

Any of these approaches works with the method above. The difference is how much of Step 2 you do yourself.

Approach	Effort	What you get	Cost
By hand (calendar, notes, sent mail)	High. You dig through sources each week, plus a daily note if you keep one	Whatever you remembered or wrote down	$0
Toggl (formerly Toggl Track), timer app	Medium. You start and stop timers or fill in timesheets during the week	Time by project and client, with billable rates on paid plans	Free plan for up to 3 users. Starter is $9 per user per month billed annually, or $12 per user per month billed monthly
Timing (automatic)	Low. You review and assign suggested entries	Apps, websites, documents and URLs on the Mac, plus AI activity summaries	Professional is $9 per month billed annually, or $11 per month billed monthly. 30-day trial
Rize (automatic)	Low. You check the AI categories	App and website time, AI categorization and AI insights	Basic is $9.99/mo billed annually, or $12.99/month billed monthly. 7-day trial, no free tier
ActivityWatch (automatic)	Medium. Some setup, and you build your own categories	App, window title and away-from-keyboard data stored on your device	Free and open source (MPL-2.0)
Dayflow	Low. You read and correct the timeline	An AI-written timeline built from screen recordings, with Markdown export	$0 (MIT), with an optional Pro plan at $20 a month

Again, Dayflow is our product. Its recordings and timeline are stored on your Mac. Screenshot analysis can run in one of three ways:

• Locally through Ollama or LM Studio, so it stays on your Mac.
• Through an AI provider you configure, under that provider’s terms.
• Through Dayflow Pro. Under the privacy policy, Pro deletes screenshots after processing and may retain activity data for sync.

The local and cloud processing page covers the trade-offs.

Data location for the others:

• Timing stores data locally unless you opt into sync.
• ActivityWatch keeps everything on your device.
• Rize hosts data on Amazon Web Services.
• Toggl is a cloud service that syncs across its apps.

Common mistakes

• Skipping the review in a bad week. Bad weeks hold the most information. Do a short version rather than none.
• Reviewing without a record. You will review the week you remember, not the one you had.
• Changing buckets every week. You lose the ability to compare one week with the next.
• Turning the review into a to-do list dump. Capture loose tasks in your task manager. The review is for deciding, not storing.
• Spending the whole slot on Step 2. If gathering takes most of the 45 minutes, keep a daily note or let software collect the record.
• Sharing AI-drafted summaries unread. Whatever tool drafts it, you are the one sending it.
• Never rereading old reviews. Once a month, skim the last four. Patterns that are invisible in one week are obvious across four. If you want to look at focus specifically, see our deep work review.

Where Dayflow is the wrong pick

Dayflow is built for one person on a Mac who wants a readable record of the day. Plenty of weekly reviews need something else.

• You bill clients and need invoices or team reports. Dayflow has no invoicing, payroll approvals or team reports. Toggl is the better fit. It offers billable rates and team reports, plus timesheet approvals and profitability reports on Premium. Timing is the Mac-only alternative, with billing status tracking and a GrandTotal invoicing integration depending on plan.
• You work on Windows or Linux, or across several machines. Dayflow is Mac only and needs macOS 14 or later. ActivityWatch is free and runs on Windows, Mac, Linux and Android. Rize covers Mac and Windows.
• You want app and window data without screen recording. ActivityWatch records app names, window titles and away-from-keyboard time on your device. Toggl states that it does not take screenshots or record keystrokes.
• You need an older version of macOS. Rize lists macOS Mojave 10.14.6 or later.
• You want the review to come with distraction blocking. Dayflow has no website blocking. Check each vendor’s feature list, or pair your tracker with a dedicated blocker.

If your review is mainly about client hours, Toggl or Timing will serve you better than Dayflow.

The bottom line

A weekly review is six steps: protect the slot, gather the record, sort it, compare it with the plan, write five lines and set three priorities. The method costs nothing and works with a calendar and a text file.

Software earns its place in Step 2:

• You already run timers for billing: Toggl gives you the numbers.
• You want your Mac to record apps and documents for you: Timing and Rize both do that. Timing keeps data local by default, and Rize keeps it in the cloud.
• You want free and cross-platform: ActivityWatch fits.
• You want a written timeline of what you did, free on a Mac: try Dayflow, and export the week as Markdown before your next review.

Frequently asked questions

How long should a weekly review take?

Plan on about 45 minutes once you have a routine. The first few reviews take longer because you are still building categories and a running document. If gathering the record eats most of the slot, keep a short daily note or use software that records your week. That way the time goes to deciding what to change, not to reconstructing what happened.

What questions should I ask in a weekly review?

Start with three. Did my top priorities get the time they needed? What took longer than expected? What unplanned work showed up, and was it worth it? Then write five lines covering what shipped, what moved, what is stuck, one lesson and the headline for next week. Finish by picking three priorities and booking calendar time for each.

Is Friday or Monday better for a weekly review?

Friday afternoon is better for accuracy because the week is still fresh, and you start Monday with a plan already set. Monday morning suits people who plan better when rested, but midweek details fade over a weekend. If you choose Monday, keep a daily note so you are not relying on memory. Doing the review every week matters more than which day you pick.

Can I do a weekly review without a time tracker?

Yes. A calendar, sent email, chat history, your task board and a few lines written at the end of each day are enough. The weak spot is work that leaves no trail, like a long debugging session or a day of reading. A daily note covers most of that. A tracker only removes the gathering work. It does not do the thinking for you.

Is there a free app that records my work week on a Mac?

ActivityWatch is free and open source under MPL-2.0. It records apps, window titles and away-from-keyboard time, and stores the data on your device. Dayflow costs $0, is MIT licensed, needs macOS 14 or later, and writes an AI timeline from screen recordings stored on your Mac. Toggl has a free plan for up to 3 users, but it relies mostly on timers you start yourself.

What is better than Dayflow for tracking billable hours?

Toggl or Timing. Dayflow has no invoicing, payroll approvals or team reports. Toggl offers billable rates on paid plans and team reports, plus timesheet approvals and profitability reports on Premium. Its Starter plan is $9 per user per month billed annually. Timing is Mac only and offers billing status tracking and a GrandTotal invoicing integration depending on plan. Its Professional plan is $9 per month billed annually.

Does Dayflow upload my screen recordings?

Full recordings and the timeline are stored on your Mac. Screenshot analysis can run on-device through Ollama or LM Studio. If you configure your own cloud provider, selected screenshots go to that provider under its terms. With Dayflow Pro, the privacy policy says screenshots are deleted after processing and activity data may be retained for sync.

Can AI write my weekly review for me?

AI can draft the summary from a record, such as a Dayflow Markdown export or a tracker's activity summary, which saves typing. It cannot decide what mattered or what to change next week, and AI summaries can miss or misinterpret work. Read every line before you share it. If you use a cloud assistant, your data goes into that conversation even if it was stored locally.

How do I do a weekly review on Windows?

The method is the same. Only the recording tool changes, because Dayflow is Mac only. ActivityWatch is free and runs on Windows, Mac, Linux and Android. Rize supports Mac and Windows, with Basic at $9.99 a month billed annually and a 7-day trial. Toggl has web, Windows and mobile apps and a free plan for up to 3 users.

What is the difference between a weekly review and a time audit?

A time audit is a one-off, detailed look at where your hours go, usually over a week or two with fine-grained categories. A weekly review is a short recurring routine that compares the week with your plan and sets the next one. Run an audit when you want to redesign your schedule. Run a review every week to keep that schedule on track.