Deep work means long, uninterrupted stretches on cognitively demanding tasks. The practical tips that matter are: schedule the block before the day starts, protect it from interruption rather than resisting interruption, batch the shallow work into its own slot, and finish at a point that’s easy to restart from. Two protected hours a day beats eight hours of fragments.
The term comes from computer scientist Cal Newport’s 2016 book Deep Work, which argues that the ability to concentrate without distraction is becoming both rarer and more valuable. The idea is popular and the specific claims in the book aren’t all backed by controlled research, so below we separate the practices with evidence behind them from the ones that are simply sensible.
Why is fragmented work so costly? #
Two findings explain most of it.
Switching costs time, and the cost grows with complexity. In classic experiments by Rubinstein, Meyer and Evans (2001), people were slower when alternating between tasks than when repeating one, and the penalty rose as the rules got more complicated. Your work is far more complex than their lab tasks, so the rebuild is bigger.
Part of your attention stays behind. Management researcher Sophie Leroy found that when people switch away from an unfinished task, their attention stays partly stuck on it, and performance on the next task suffers. She called it attention residue. It’s why a two-minute email inside an hour of writing costs much more than two minutes.
There’s a complication worth knowing. In a 2008 experiment by Gloria Mark, Daniela Gudith and Ulrich Klocke, interrupted workers finished faster with no drop in quality, but reported more stress, frustration and effort. They compensated by rushing. So interruption doesn’t always show up as less output. Sometimes it shows up as a worse day. We go through the detail in the real cost of task switching.
10 deep work tips #
1. Put the block in the calendar before the day starts #
A deep work block that isn’t scheduled becomes whatever is left after everything else, which is nothing. Pick the slot the night before and treat it as an appointment with a start time.
2. Give the block one named output #
Not “work on the paper”. “Draft the results section.” A named output tells you when you’re done and makes drifting obvious, because you can see you’re not doing it.
3. Protect the block instead of resisting interruption #
Close the mail client, quit the chat app, and tell people when you’ll be back. Every interruption you have to decline still costs you the noticing. The aim is a block where the interruptions never arrive.
4. Start with the same short ritual #
Same desk, same drink, same two minutes of reading yesterday’s last paragraph. The ritual isn’t magic, it’s a cue. Habits form through repetition in a stable context, which is what Lally and colleagues tracked when they followed 96 people repeating a daily behaviour until it felt automatic.
5. Batch the shallow work #
Email, messages, admin and small requests go in one or two defined slots a day. Thirty scattered switches cost far more than one block of switching, and most “urgent” requests survive four hours.
6. Work long enough to get past the warm-up #
Demanding work has a start-up cost. If your block is 25 minutes, a good share of it goes on reloading context. For writing, coding and hard problem sets, 60 to 90 minutes is usually the floor. The 52/17 rule is one popular version of a long block with a real break, though its evidence base is thinner than people assume.
7. Stop at a restartable point #
Finish the paragraph, the function or the proof step before you break, and write one line about what comes next. That line turns tomorrow’s warm-up from ten minutes into thirty seconds. Some writers deliberately stop mid-sentence for the same reason; try both.
8. Keep a distraction list #
When a thought arrives during the block (“book the dentist”, “check that figure”), write it down and keep going. The thought stops nagging without becoming a switch. Deal with the list in the shallow slot.
9. Do one block, not four #
Two protected hours a day, every day, is a serious amount of work. Four blocks scheduled on ambition and then abandoned teach you that the schedule is fiction. Start with one block you will definitely do.
10. Measure blocks completed, not hours at the desk #
Hours at a desk is the metric that lies. Blocks completed without leaving is the one that tracks the thing you care about.
How do you stop your phone breaking a deep work block? #
The phone is the interruption you carry, and it’s immune to closing your mail client. If you can, the answer is distance: another room, or at minimum out of arm’s reach with notifications off at the app level rather than on silent. Even unanswered notifications disrupt attention at a size comparable to actually using the phone.
If your phone has to be on the desk, make leaving it cost something. Work Capy is a focus timer built around that idea: you name the task, start a session of up to three hours, and put the phone face down. Leave the app for more than about five seconds and you get one warning; leave again and the session is forfeit, earning nothing and adding no minutes to your total. Locking the screen never counts as leaving, and pausing is free, so a genuine interruption costs nothing if you pause first.
It also has a Desk Mode that keeps the screen awake and dimmed, which turns the phone into a countdown you glance at rather than an object you pick up. What it can’t do is stop you switching tabs on your laptop. That’s still on you.
Does deep work suit everyone? #
No, and it’s worth being honest about the trade-offs.
- Jobs with real-time duties. Support, clinical work, teaching and operations roles don’t have four-hour holes in them. A protected 45 minutes is a better target than a fantasy morning.
- Work that depends on other people. If three colleagues are blocked waiting on your answer, disappearing for a morning moves the cost onto them rather than removing it.
- People who find long blocks aversive. If a 90-minute block is why you never start, shorter rounds are better. Task aversiveness is one of the strongest predictors of procrastination in Piers Steel’s 2007 meta-analysis, so a method you dread is a method you’ll avoid. Try how to stop procrastinating first.
- Collaborative and generative work. Some problems get solved faster by talking to someone than by staring at them alone for two hours.
The honest summary is that deep work is a strong default for individual cognitive work, and a bad default for everything else. Choosing between long blocks and a scheduled day is the subject of time blocking vs Pomodoro.
Frequently asked questions #
How many hours of deep work can you do in a day? #
Most people manage two to four hours of genuinely demanding concentration, and Newport suggests something in that range for people practised at it. Beginners should expect less. The rest of a working day goes on shallow tasks, which is normal rather than a failure.
Is deep work scientifically proven? #
The underlying pieces have research behind them: switching costs, attention residue, and the value of planned breaks. The specific package Newport recommends hasn’t been tested as a whole in controlled trials, so treat it as a well-argued framework rather than a proven protocol.
What’s the difference between deep work and the Pomodoro Technique? #
Length and intent. Pomodoro uses short fixed rounds with frequent breaks and is good for starting dull or daunting tasks. Deep work aims for long uninterrupted stretches on demanding ones. Many people use Pomodoro to get moving and longer blocks once they’re in.
How do I do deep work in an open-plan office? #
Book a room or use the quietest hours, usually early morning. Failing that, signal unavailability clearly, put headphones on, close every message app, and keep blocks to 45 or 50 minutes so they fit between meetings. Protecting one short block reliably beats planning a long one that gets eaten.