Guide
A decision log is a dated record of each significant choice: what you decided, what you knew, what you rejected and when you will check the result. A plain document or spreadsheet is enough. On a Mac, Dayflow helps rebuild the context around a decision, Timing suits people who also need billable-hour reports, and ActivityWatch is the free option for Windows or Linux.
A decision log is a running, dated record of the choices you or your team made, with the reasoning behind each one, the options you turned down and a date to check whether it worked. Each entry is written at the moment of deciding, not reconstructed afterward.
Most people get it wrong in one of two ways. Some treat it like meeting minutes and record everything that was said. Others only write it up after something goes badly, when hindsight has already rewritten what they knew and why they chose what they did. The point of the log is to freeze your thinking before the outcome arrives.
Disclosure: we make Dayflow, a Mac app that appears in the tools section near the end of this guide.
A decision log is a single place where decisions get written down in a consistent shape. It can belong to one person, a project or a whole team. What makes it a log rather than a pile of notes is that every entry has the same fields, so you can scan, sort and compare entries months later.
A complete entry usually has these fields:
| Field | What goes in it |
|---|---|
| Date | The day the decision was made |
| Decision | One sentence, stated plainly |
| Context | What you knew and what constraints applied, as of that day |
| Options rejected | The alternatives you seriously considered and why they lost |
| Prediction | What you expect to happen, with a rough confidence level |
| Owner | Who made the call or is accountable for it |
| Review date | When you will check the result |
| Outcome | Left blank until the review |
Not every choice deserves an entry. The decisions worth logging are the ones that are costly or slow to reverse, the ones someone will ask you to explain later, and the ones where you are genuinely unsure. Picking a font for an internal slide does not qualify. Switching a client from hourly to fixed pricing does.
Some people keep a private decision log for their own career and money choices. Teams keep shared ones so that new members can see why things are the way they are without asking around. Software teams often keep a specialized version called an architecture decision record, stored as short files alongside the code.
The decision log overlaps with three other records, and mixing them up is how logs get bloated or abandoned.
| Record | What it captures | When it is written | Question it answers |
|---|---|---|---|
| Decision log | Choices, reasoning, predictions | At the moment of deciding | Why did we do this, and did it work? |
| Meeting notes | Discussion from one meeting | During or right after the meeting | What was said? |
| Work journal | What you did and thought each day | End of each day | What did I work on? |
| Changelog | What changed in a product or system | When the change ships | What is different now? |
Meeting notes are organized by meeting. Decisions get buried among discussion points, and decisions made outside meetings (in chat, in a doc comment, alone at your desk) never appear at all. A decision log pulls the choices out into one list regardless of where they happened.
A work journal is broader and more personal. It records activity, progress and reflections day by day. A decision log is narrower and more structured. Many people keep both, with the journal feeding the log: you notice in your journal that you made a call, then write it up properly. Our guide on what a work journal is covers that habit in detail.
A changelog records results, not reasons. It tells you the pricing page changed on a given date. It does not tell you that you raised prices because two competitors had, that you expected some churn and that you planned to check in eight weeks.
Memory is unreliable in a specific direction. Once you know how something turned out, you tend to remember your earlier thinking as more certain and more aligned with the result than it was. A bad outcome makes the decision feel obviously wrong in retrospect. A good outcome makes it feel obviously smart. Neither feeling tells you much about whether the reasoning was sound.
A decision log fixes this by recording your thinking before the result exists. When you review an entry, you are comparing two things you wrote down: what you expected and what happened. That comparison is where you learn.
It also does several practical jobs:
There are many templates, but they boil down to the same loop. Here is a seven-step version we call Capture, Predict, Review.
The whole entry should take five to ten minutes. If it takes longer, the template is too heavy.
This is a made-up example to show the shape of a real entry.
Priya is a freelance product designer. One of her clients has been on hourly billing for a year, and every invoice triggers a back-and-forth about line items. She decides to propose a fixed monthly retainer instead. Here is her entry:
At the end of June, Priya fills in the outcome:
Without the log, Priya would probably remember this as “retainers lower my rate” and avoid them. With it, she can see the real lesson was about scope definition, not about retainers. If she wanted to check the effective rate more carefully, a tool like the real hourly rate calculator would give her the number.
When every small call gets an entry, the log becomes noise and nobody reads it. Stick to your trigger rule. A log with fifteen entries a quarter that all get reviewed is worth more than one with two hundred that never do.
An entry written a week later is already contaminated by what happened in that week. If you cannot write the full entry on the day, write the one-sentence decision and the prediction immediately, then fill in the rest within a day.
Without the alternatives, a review only tells you whether the chosen path worked. It cannot tell you whether another path would have worked better, or whether you even considered the right ones.
“Should improve things” cannot be wrong, so it cannot teach you anything. Name a measurable result and a rough probability.
The review is the whole point. A date in a spreadsheet column that nobody checks is the same as no date. Put it in a calendar.
For a team, a decision log that lives in one person’s private notes does little. Put it where everyone affected can read it, and link to entries when the decision comes up again.
If reviews turn into finding someone to blame for bad outcomes, people stop writing honest predictions and start writing defensive ones. Grade the reasoning, not the person.
Most decision logs die in the second or third month. The fix is to attach the log to habits you already have rather than treating it as a new one.
The hardest part is the context field. On the day you decide, you rarely think to write down which documents you read, which conversations shaped your view, or how long you spent weighing options. Three months later, that context is gone. This is where activity records help: if you have a history of what you were working on that week, you can rebuild the context instead of guessing. Our piece on how to remember what you worked on covers that problem more broadly.
None of the tools below is a decision log. The log itself belongs in a document or spreadsheet you control. What these tools do is record your working activity automatically, which gives you a factual record of the context around a decision: what you had open, what you were researching, and how much time the decision took. Any automatic time tracker for Mac can play this supporting role; three are covered here.
Timing is a paid Mac app that records time across apps, websites and documents, down to specific URLs and Mail threads, and adds AI activity summaries on all plans. See our Timing comparison. ActivityWatch is free, open source under MPL-2.0 and runs on Windows, Mac, Linux and Android, storing data on your own device. See our ActivityWatch comparison. Tool details come from each vendor’s own documentation as of October 4, 2026, and Dayflow’s from its 2.2.0 release and published policies.
| Tool | What it does for this | Price |
|---|---|---|
| Dayflow | Records the screen at one frame every 10 seconds and has AI write a readable timeline of your day, so you can look back at what you were doing when a decision was made. Export a date range as Markdown to paste context into an entry. | $0 (MIT licensed); optional Pro plan at $20 a month |
| Timing | Records time across apps, websites, documents, URLs and Mail threads, with AI activity summaries and AI-powered suggestions for time entries. Useful when decisions also need to be tied to billable hours. | Professional from $9 per month billed annually, or $11 per month billed monthly; no free tier; 30-day trial, no credit card required |
| ActivityWatch | Records active app, window title and away-from-keyboard time; browser extensions add tab titles and URLs. A raw, local activity history you can query through its local API. No AI summaries listed. | Free |
Dayflow is our product. It fits when you are on a Mac and want a plain-language account of what you were doing around a decision without starting timers, and the export work history workflow shows how to pull a date range out as Markdown; with Local AI, screenshot analysis stays on your Mac, while your own cloud provider or Dayflow Pro processes selected screenshots off-device. It does not fit if you need invoicing, payroll approvals, team reports or website blocking, and its AI summaries can miss or misinterpret work, so check anything you copy into a decision entry.
If you are on Windows or Linux, or you want one tracker across a Mac and an Android phone, ActivityWatch wins. Dayflow runs only on macOS 14 or later. ActivityWatch is free, open source and keeps data on your device, though device sync is still listed as work in progress.
If your decisions are tied to client billing, Timing wins. It offers billing status tracking, report rounding and, depending on plan, GrandTotal invoicing integration and two-way Clio, ClickUp and Linear sync. Dayflow has no invoicing or billing features at all. If you need a shared view for a team, Timing’s Connect plan adds shared projects, while Dayflow has no team reports.
A decision log is a short, structured record of the choices that matter, written on the day you make them. The fields are simple: the decision, the context, the options you rejected, a specific prediction and a review date. The value comes almost entirely from two habits: writing the entry before you know how things turn out, and actually reviewing it when the date comes.
Start with a spreadsheet and a clear trigger rule. Log fewer decisions than you think you should. Put review dates in your calendar. Grade the reasoning separately from the result. After a few months, you will have something most people never get: an honest record of how well you decide, not just how lucky you were.
Each entry needs the date, the decision in one sentence, the context as you understood it that day, the options you rejected, your reasoning, a specific prediction with a confidence level, an owner and a review date. Leave an outcome field blank to fill in at the review. Anything beyond that is optional. Most people keep entries under a hundred words so the habit survives a busy week.
Meeting minutes record what was said in one meeting, including discussion that led nowhere. A decision log records only the choices, wherever they were made: in meetings, in chat, or alone at your desk. Minutes are filed by meeting date and rarely reopened. A decision log is a single running list you go back to, with a review date on each entry so outcomes get checked.
An architecture decision record, or ADR, is a decision log entry for software design, usually stored as a short file in the code repository next to the code it affects. The structure is similar: context, decision, consequences. A general decision log covers any kind of choice, from pricing to hiring to which client to drop, and adds a prediction and a review date that many ADR templates leave out.
Review each entry on the date you set when you wrote it, which is usually somewhere between two weeks and six months out depending on how fast results show up. A fifteen-minute weekly pass, attached to a ritual you already have like a Friday status report, is enough to catch entries that are due. Reviews that never happen are the most common reason a decision log stops being useful.
Yes, and a spreadsheet works well because you can sort by review date and filter by outcome. Use one row per decision and one column per field. A notes app or a single running document also works. The tool matters far less than writing the entry on the day you decide and actually reviewing it later. Pick whatever you already open every day.
No. A time tracker records what you were doing, such as apps, windows and documents, but it cannot record why you chose one option over another or what you expected to happen. What a tracker can do is help rebuild the context around a decision: which documents you read, what you were working on that week and how long the evaluation took. The reasoning still has to be written by you.
No. Dayflow runs only on macOS 14 or later, so it is not an option on Windows or Linux. ActivityWatch is the better pick there: it is free, open source under MPL-2.0, runs on Windows, Mac, Linux and Android, and stores data on your own device. It records active apps, window titles and, with browser extensions, tab titles and URLs, which you can use as context when writing entries.
If billing is part of the job, Timing is the stronger choice. It offers billing status tracking, report rounding and, depending on plan, GrandTotal invoicing integration and Clio sync, starting at $9 per month billed annually or $11 billed monthly. Dayflow has no invoicing. Dayflow fits better if you mainly want a free, readable AI timeline of your day to look back on when writing up what you decided and why.
Log choices that are expensive or slow to reverse, choices someone will later ask you to explain, and choices where you are unsure and want to learn from the result. Skip routine calls you make the same way every time. A useful filter: if you would struggle in three months to explain why you did this, it belongs in the log.
Dayflow is $0 and open source. Full recordings stay on your Mac, with local AI available for on-device analysis.
Download for MacOS macOS 14+ · MIT licensed · no account required