A lone climber silhouetted against a mountain at sunset — single focus, sustained.

Photo by Sandro Tedeschini on Pexels

deep-workautomationhow-to

How to Track Deep Work on Mac (Without Manual Timers)

Manual Pomodoro timers work until they don't. You forget to start one, or you stop them mid-session. Automatic tracking fixes both failure modes — here's how to set it up on a Mac.

Mike6 min read

How to Track Deep Work on Mac (Without Manual Timers)

The usual advice for tracking deep work is a manual timer: a Pomodoro app, a stopwatch, or a hotkey to toggle a "focus session." This works for about two weeks and then it stops, for the same reason most journals stop. You forget to start the timer. You forget to stop it after the session. You start it, then switch to Slack for five minutes, then come back and realize the whole "session" was wrecked but the timer kept running.

Automatic deep work tracking solves both failure modes. The tracker runs continuously, detects when you're in productive work, and rolls up those stretches into sessions after the fact. You don't toggle anything. The data is there at the end of the day whether you remembered or not.

Here's how to get this working on a Mac with minimal setup, what "deep work" actually means in the tracker's definition, and what to do with the results.

What "deep work" means, operationally

For automatic detection to work, the tracker needs a mechanical definition of deep work. The one that holds up well across professions:

A deep work session is a continuous stretch of at least N minutes in an app or site you've tagged as productive, with fewer than K context switches and no idle gap longer than I minutes.

In Focus Meter's default config, that's N=25, K=2, I=3. Tweak to taste, but the defaults are sensible for most knowledge workers. The 25-minute floor matches the Pomodoro literature; the 2-switch tolerance lets you alt-tab to a terminal once without wrecking the session; the 3-minute idle cap pauses the count when you step away.

This gives you two kinds of numbers:

  • Total deep work time per day — the sum of all qualifying sessions.
  • Deep work session count and length distribution — how many sessions, how long each one was.

Both are useful. The first is the headline. The second is what tells you whether your day was "one long flow session + lots of fragments" vs. "three solid 45-min stretches" — a real difference in how the day felt that totals alone hide.

The setup (about five minutes)

Assuming you're starting with nothing:

1. Install a tracker with automatic session detection. Focus Meter is the Mac-native option designed for this. Alternatives: Timing (heavier, more for billing), RescueTime (cloud-based). Install it and let it run.

2. Tag your productive apps. On day one, the only thing the tracker knows about your apps is their names. It needs you to tag which are productive, which are distracting, and which are neutral. Start with your top 10-15 apps — you can tag more later as they appear.

A rough first pass for a developer:

  • Productive: your editor, terminal, browser (conditionally on URL), Linear/Jira, Figma if you design.
  • Neutral: Finder, Spotlight, password managers.
  • Distracting: Slack, Twitter, YouTube (unless tutorials), Reddit.

A rough first pass for a writer:

  • Productive: your writing app, research tabs, note-taking app.
  • Neutral: email (in moderation), calendar.
  • Distracting: social media, news sites.

Don't stress about getting this perfect. The first pass takes 5 minutes and you can refine it as you see results.

3. For browser-heavy work, enable URL tracking. This is the critical step for people whose work lives on the web. Without it, "Chrome" is one undifferentiated blob and the tracker can't tell docs.google.com apart from reddit.com. With it, you can tag individual domains, which is what makes the productive / distracting classification accurate.

Focus Meter → Settings → Privacy → turn on Browser URL Tracking, grant Automation permission per browser.

4. Work normally for a day. No need to start sessions, press hotkeys, or think about it. Live your day.

5. Open Reports the next morning. Look at the Deep Work section. Total time, session count, session length distribution.

What the first week of data will show

Three patterns show up reliably in the first week.

Your headline deep-work number will be lower than you expected. Most people estimate they do 4-5 hours of deep work on a good day. Measured honestly — with idle time excluded, Slack-breaks counted as switches, and meetings not counted as deep work — the actual number is usually 2 to 3.5 hours for a good day, and under 1 hour for a meeting-heavy day. This isn't failure; it's reality. 3 hours of actual deep work is a productive professional day.

You'll have at least one day where the data shocks you. Usually it's a day that felt busy but produced nothing, and the tracker shows 40 minutes of deep work with 120 context switches. The feeling of being "slammed" turns out to be mostly Slack. This day is what makes the tracking worth it.

Your session length distribution will cluster bimodally. Most people have lots of short sessions (under 20 min) plus occasional longer ones (45-75 min). The long ones are almost always the output-producing sessions. The short ones are mostly "I tried to work, then got pulled into a message." The question the data poses is how to trade short sessions for long ones.

What to do with the data week two onward

Don't try to maximize the number. The goal isn't 8 hours of deep work per day — that's not a realistic target and chasing it leads to weird workarounds (leaving your editor open on lunch). The goal is a trend line that's slightly up, week over week, with at least one 45-min+ session most days.

Three specific experiments worth running, each for about a week:

  • No Slack in the first 90 minutes. Measure the impact on morning deep-work time. For most people, this alone adds 30-60 minutes to the daily total.
  • Notifications off entirely. Measure the impact on switches per hour. This is usually the biggest single lever.
  • One "deep work" block on the calendar. 90 minutes, no meetings allowed. Measure whether you actually use it; measure the session length you hit. Most people who block the time discover the constraint wasn't time, it was focus habits, and the block helps but doesn't fix it.

Why this beats manual timers

Manual timers require discipline about an action (starting the timer) that itself is often the thing your focus is struggling with. When you're about to slide into an hour of Twitter at 10am, you're not going to remember to start a timer that tells you that's what you're doing. Automatic tracking doesn't require that discipline. The number is there at the end of the day.

The other benefit: you can look at data from three months ago. Manual timer data is whatever you happened to remember to log. Automatic data is a continuous record of how your attention actually behaved, independent of what you felt like reporting.