How to measure deep work without using a stopwatch
· 6 min read
Measure deep work by the time your main work app held the front of the screen, not by a timer you remember to start. A timer measures intention; the front app measures what happened. Over a week, the main app’s daily total is a fair upper bound on deep work, and a few honest adjustments turn it into a number you can trust and compare.
Why the stopwatch fails for deep work
Deep work has one defining feature: you are not thinking about anything else, including whether you are doing deep work. A stopwatch requires you to think about exactly that, at exactly the moment you should not be.
So the timer goes wrong in predictable ways. You forget to start it and the best hour of the day goes unrecorded. You start it, get interrupted, forget to stop it, and a ninety-minute “session” contains a twenty-minute conversation. You remember halfway through and guess the start time. By Thursday the numbers are half fiction and you stop trusting them, and by the next week you have stopped using the timer at all.
The deeper problem is that the timer measures the decision to focus, not the focus. You pressed start. Whether the next hour was deep is a separate question the timer cannot answer.
A definition you can actually measure
“Deep work” is usually defined by what it feels like, which is why it is hard to count. Swap in a definition based on what the screen shows.
Deep work is time in which the app that produces your real output was in front, in stretches long enough to matter, with few switches away. For a developer that app is the editor. For a designer it is the design tool. For a writer it is the writing app. For an analyst it is the notebook or the spreadsheet.
This definition is imperfect on purpose. Reading documentation in a browser can be deep. Sitting in the editor rereading the same function can be nothing. But it has two properties the felt definition lacks: it can be recorded without your participation, and you cannot easily fool yourself about it afterward. A week of front-app records will tell you something true. A week of felt deep work will tell you how you felt on Friday.
The method: a week of front-app totals
- Name your deep work app. One, or at most two. Write it down before you start so you do not change it to flatter the result.
- Guess your daily deep work in hours. Write that down too.
- For one normal week, get a per-app record of each day. You can keep one by hand (note the time whenever you switch), use Screen Time in System Settings (it reports per app but is built for limiting use rather than reviewing work), or use a menu bar tracker that records the front app automatically.
- Each evening, write down the deep work app’s total for the day.
- On Friday, add up the five days and divide by five. That is your daily average, and it is your upper bound: the most deep work you could have done this week.
- Subtract what you know was not deep. The forty minutes in the editor during a screen-share. The hour you were in the writing app but actually in a meeting on the phone. Be honest, be rough, and write the adjusted figure next to the raw one.
- Compare the adjusted figure with your guess from step 2.
The gap between steps 2 and 7 is the finding. Most people guess high. Why your estimate of your own workday is wrong by hours covers why.
Reading the number as an upper bound
The method gives you a ceiling, not a measurement, and it is worth being precise about what that means.
The raw total says: the deep work app was in front for this long. Everything that counts as deep work happened inside that time. Nothing outside it does. So if the editor total is two hours and ten minutes, deep work was at most two hours and ten minutes, and probably less.
That sounds like a weakness. It is the useful part. A ceiling of two hours in an eight-hour day tells you something no timer would: that six hours went elsewhere, regardless of how deep the two were. If the ceiling is low, the question is not how to focus better inside the editor. It is where the other six hours went, and Where does the workday actually go? How to find out on a Mac takes that up.
If the ceiling is high but the work still felt shallow, the problem is fragmentation: the editor had four hours, but in twenty-minute pieces. The front-app total cannot show that directly. You can get at it roughly by counting how many other apps got meaningful time on the same day; a day with the editor at four hours and eight other apps at fifteen minutes each was a day of switching.
Using a receipt instead of a log
Steps 3 and 4 are where people quit, because they require doing something every day that has nothing to do with the work.
Punchcard takes those two steps away. It sits in the menu bar and notices which app is in front during the day, by app name only. It never sees window titles, document names, or what is on screen, and it asks for no macOS permissions. At the closing time you set, it prints a receipt: one line per app with the time, a day total, a stamp. Your deep work app’s total is one line. The week roll on Sunday is step 5 done for you.
It does not know what deep work is. There is no “focus” category, no way to mark a session, no switch count, and no timer to start. It gives you the honest ceiling and leaves the adjustment in step 6 to you, which is where it belongs, because only you know that the editor hour on Wednesday was a screen-share. If you want a tool that measures focus sessions on purpose, with a start and a stop, this is not that; it is the opposite approach.
What this still cannot tell you
No front-app method knows whether the work was good. An afternoon of deep, productive work and an afternoon of deep, productive procrastination inside the same app look identical.
It cannot see deep work that happens away from the screen. Thinking on a walk, sketching on paper, a whiteboard session: all invisible. If those are a large part of your real work, add them by hand and do not feel the number is lying to you.
And it cannot distinguish reading from writing. Two hours in the editor reading someone else’s code and two hours writing your own are the same line on the receipt.
These limits are why the number is a ceiling. Treat it as one and it is honest. Treat it as a score and it will mislead you.
Questions
Is two hours of deep work a day low?
It is common. Whether it is low depends on the job. What the measurement gives you is the number; whether to change it is a decision about your week, not a verdict from the tool.
Can I measure deep work inside a browser?
Only as browser time. Without URLs, which a no-permissions tracker cannot see and a privacy-respecting one should not want, the browser is one line. If your deep work lives in a browser app, treat the browser as the deep work app for the week and accept that it includes everything else the browser does.
Why not just use a focus timer with a good habit?
If you have the habit, keep it. Most people do not, and the timer’s numbers decay within a week. The front-app method needs no habit, which is the whole reason it survives.
Does Punchcard count switches?
No. It prints time per app and a day total. Fragmentation has to be read indirectly, from how many apps got real time on the same day.