How Long Should a Focus Session Be? A Framework Beyond the 25-Minute Default

Twenty-five minutes is not a law of cognition. Francesco Cirillo picked it in the 1980s because it fit on a kitchen timer shaped like a tomato, and it stuck because it’s a reasonable default for a lot of shallow work. But if you’ve ever hit a wall at minute 22 of a task that clearly needed another twenty minutes to land, or felt restless five minutes into a session that was too long for what you were actually doing, the number was wrong for that moment — not the technique. This is a framework for picking the right length instead of inheriting one.

Why 25 minutes became the default

The Pomodoro Technique’s genius was never the specific number — it was the shape: bounded effort, followed by bounded rest, repeated. Twenty-five minutes works as a starting point because it’s short enough that almost any task can survive an interruption at that mark, and long enough to produce a meaningful chunk of output. That makes it a safe default for people who’ve never timed their own work before. It was never claimed to be optimal for everyone, every task, every time of day — that part got lost as the number calcified into “the Pomodoro length” instead of “a reasonable starting length.”

The three variables that actually set the number

Session length isn’t one setting — it’s the output of three things that change by task and by day:

• Task type. Shallow, mechanical work (clearing an inbox, tagging expenses, updating a spreadsheet) rarely benefits from long blocks — you’re not building up state in your head that a break would destroy, so 15–20 minutes is often enough and longer sessions just invite fidgeting. Deep work — writing, debugging, design — has ramp-up cost. If it takes you eight minutes to load the problem into working memory, a 25-minute session gives you maybe 17 minutes of actual thinking before the timer interrupts you mid-thought. That work wants 45–90 minute blocks.
• Energy, not just time of day. Most advice says “mornings are for deep work,” but the honest version is narrower: you have a window, somewhere in your day, where your attention is hardest to knock loose, and it isn’t the same hour for everyone. Track two or three sessions at different times and note not how much you finished, but how many times your attention drifted. That’s the more reliable signal.
• Interruption risk. A block with no risk of being pulled away — phone off, door closed — can run long, because the only thing ending it is you. A block where a Slack message or a kid walking in is genuinely likely should stay short, because a 90-minute session that gets cut at minute 12 is worse than three honest 20-minute ones. Pick the length your environment can actually protect, not the length your ambition wants.

None of these are things you can look up. They’re specific to the task in front of you, your body that day, and the room you’re in right now — which is exactly why a single fixed number was always going to be wrong some of the time.

A simple way to find your number

You don’t need a spreadsheet for this, just a week of noticing:

1. Run three different lengths on comparable tasks — a 20-minute block, a 45-minute block, a 90-minute block — each on work of similar depth. Don’t optimize yet, just observe.
2. Rate attention, not output. After each session, ask one question: did my focus hold to the end, break early, or want to keep going past the buzzer? Output varies too much day to day to be a reliable signal on its own; sustained attention is the thing you’re actually measuring.
3. Notice where sessions want to keep going. That’s a stronger signal than a session that drags — it usually means the block was too short for the task, not that you have unlimited focus.
4. Set defaults per task type, not per day. You’ll likely land on something like: 15–20 minutes for admin work, 25–35 for focused execution, 50–90 for deep creative or technical work. That’s three defaults, not one number to defend.

This is close to what shows up in the research on ultradian rhythms — the idea that attention naturally cycles in roughly 90-minute waves with dips in between — but you don’t need to take that on faith. The three-length test above will tell you your own pattern faster than reading about someone else’s.

Building it into a routine instead of re-deciding daily

The point isn’t to pick a length every morning — that’s its own kind of friction. It’s to have two or three named session shapes ready, matched to task types you already recognize, so starting a session is a one-tap decision instead of a negotiation with yourself. That’s the model behind Mintly’s custom session lengths and named rituals: a Quick Sprint for the 15-minute admin block, a Deep Work session for the 45-minute one, without forcing either into a 25-minute shape that fits neither.

The 25-minute Pomodoro isn’t wrong. It’s a good guess for someone who hasn’t measured yet. Once you have — even just a week of noticing where your attention actually holds — you have something better than a guess: your own number, for your own kind of work.