Deep Work: The Hidden Cost of a Fragmented Day — and Where the Advice Breaks
Glance at a chat notification, answer it, glance back at the document you were writing. The switch felt free — a few seconds, no big deal. It wasn’t. A slice of your attention stayed stuck on the message, and it will drag on the next paragraph you try to write. That leftover pull has a name in the research literature — attention residue — and it is the reason a day sliced into fragments produces so much less than the hours on the clock would predict.
The cost is not hypothetical. In four experiments on task-switching, University of Michigan psychologists estimated that toggling between tasks can eat as much as 40 percent of someone’s productive time, and that the penalty grows as the work gets more complex — exactly the profile of the demanding thinking most knowledge jobs are paid for. A recent analysis of workplace telemetry found the average employee is now interrupted every two minutes during core hours — roughly 275 times a day. Run the two findings together and the modern default starts to look less like “work occasionally interrupted” and more like “interruption occasionally interrupted by work.”
Deep work is the term computer scientist Cal Newport gave to the opposite of that default. In his 2016 book he defines it as:
Professional activities performed in a state of distraction-free concentration that push your cognitive capabilities to their limit. These efforts create new value, improve your skill, and are hard to replicate.
Every clause carries weight. Distraction-free rules out the tab-flicking baseline. Push your cognitive capabilities to their limit rules out coasting. Hard to replicate is the economic point — the output of a mind at full stretch is precisely what can’t be copied or automated cheaply. This piece is about that payoff, but also about its price and its limits: what fragmentation actually costs, why the advice works when it works, and the situations where the standard deep-work playbook quietly falls apart.
Deep work vs. shallow work
Newport pairs deep work with its opposite, shallow work: non-cognitively-demanding, logistical tasks, often done while distracted, that create little new value and are easy to replicate. Routine email, a status meeting, formatting a deck, reconciling a calendar — none of it is worthless, but none of it stretches you, and all of it is easy for anyone (or any tool) to do.
The distinction matters because shallow work expands to fill whatever room you give it. It feels productive: you end the afternoon tired, pointing at a cleared inbox, while the report or the strategy or the code — the thing that actually moves your career — never got a genuine block of attention.
| Deep work | Shallow work | |
|---|---|---|
| Cognitive demand | High — pushes your limits | Low — logistical, routine |
| Attention state | Distraction-free concentration | Usually done while distracted |
| Value created | New, hard to replicate | Marginal, easy to replicate |
| Examples | Writing, coding, analysis, strategy, design | Email triage, status meetings, scheduling, formatting |
| How it scales | Compounds your skill over time | Fills whatever time you give it |
Newport’s practical rule is not to eliminate shallow work — impossible — but to be honest about how much your role truly requires, cap it, and defend the rest of your time for depth. As the next section shows, the reason to cap it is not moralistic. It is that every uncapped switch is quietly taxing the deep work you have left.
What a single switch actually costs
The intuition that “I can just jump back in” is where the money leaks. Three well-documented lines of research say otherwise, and a fourth dataset shows how relentless the switching has become.
Attention residue. The clearest evidence is organizational psychologist Sophie Leroy’s 2009 paper Why Is It So Hard to Do My Work?, published in Organizational Behavior and Human Decision Processes. Leroy found that when you switch tasks, part of your attention stays with the first task — a residue — and that leftover pull measurably degrades performance on the next one. The effect was worst when the first task was left unfinished under time pressure. So the Slack ping that yanks you off a report doesn’t just cost the seconds you spend reading it; you come back carrying residue, and the deficit lingers rather than clearing in a moment.
The switch itself is expensive. In their 2001 Journal of Experimental Psychology paper Executive Control of Cognitive Processes in Task Switching, Joshua Rubinstein, David Meyer, and Jeffrey Evans had people alternate between tasks across four experiments. Every switch imposed a time cost, and that cost grew as tasks became more complex and less familiar. Individually each switch feels like nothing — a fraction of a second — which is exactly why the losses stay invisible while they compound. Meyer’s own estimate of the aggregate drag, summarized by the American Psychological Association, is the 40-percent figure above.
Recovery is slow — with a caveat worth stating honestly. You have probably seen the claim that it takes “23 minutes and 15 seconds” to refocus after an interruption. It’s worth being precise about where that number comes from, because most articles aren’t: it traces not to a peer-reviewed paper but to a 2006 Gallup interview with informatics researcher Gloria Mark, who also noted that people typically handle about two intervening tasks before returning to the original one. Mark’s published field studies do establish that refocusing is slow and that the working day is badly fragmented — her tracking of screen-switching shows average time on a single screen collapsing from about 2.5 minutes in 2004 to roughly 47 seconds — but the exact stopwatch figure is best treated as a vivid illustration, not a physical constant. The direction is solid; the decimal places are not.
The volume is new. What Newport described in 2016 has since intensified. Microsoft’s 2025 Breaking down the infinite workday report, built on anonymized Microsoft 365 telemetry, found the average worker now receives 117 emails and 153 Teams messages a day, that half of all meetings land in the 9–11 a.m. and 1–3 p.m. windows when focus is highest, and that 48 percent of employees describe their work as chaotic and fragmented. The fragmentation isn’t a personal failing to be willpowered away; it’s the ambient condition of the tools.
| The switching problem | The number | Source |
|---|---|---|
| Productive time lost to task-switching | up to ~40% | APA / Meyer |
| Interruptions during core work hours | every ~2 min (~275/day) | Microsoft 2025 |
| Avg. time on one screen before switching | ~2.5 min (2004) → ~47 sec | Gloria Mark |
| Meetings falling in peak-focus windows | ~50% | Microsoft 2025 |
Read together, the research says the same thing four ways: attention has inertia, moving it is expensive, moving it is slow to undo, and the workplace now moves it constantly. Deep work is simply the deliberate refusal to move it.
Fitting depth into a real calendar
Knowing you should concentrate is easy; getting it onto a real calendar is the hard part, and this is where deep-work advice usually gets over-prescriptive. Newport offers four “philosophies” for scheduling depth, and the honest framing is that they are a menu matched to constraints, not a ladder you climb. The monastic approach strips away nearly all shallow obligations for long unbroken stretches — realistic mostly for people whose value comes from one clearly defined pursuit, like some writers or theorists. The bimodal approach alternates multi-day deep stretches with periods open to collaboration. The rhythmic approach installs a fixed block at the same time every day and lets habit do the enforcing — Newport’s recommended default because it removes the daily negotiation about whether to do deep work. The journalistic approach grabs focus in whatever windows appear, which sounds flexible but is actually the hardest, since it demands dropping into concentration on command — a skill that only develops after the more structured modes.
Behind all four is a piece of arithmetic Newport borrows from research on high performers and states on his site as Work Accomplished = Time Spent × Intensity:
High-Quality Work Produced = (Time Spent) × (Intensity of Focus)
The operative word is multiplication. Intensity scales hours rather than adding to them, which is why two focused hours can beat a distracted eight, and why simply working longer at low intensity barely moves output — you’re multiplying by a number close to zero. It’s also the argument for capping shallow work rather than heroically enduring it: fragmentation doesn’t subtract from your day, it multiplies the whole thing down.
One habit worth keeping regardless of philosophy is a shutdown ritual — a deliberate end-of-day routine where you review open commitments, capture anything unfinished into a trusted place, and formally declare the day closed. It’s not a flourish; it’s a direct countermeasure to Leroy’s residue. Unfinished tasks generate exactly the lingering pull that degrades the evening’s rest, and rest is what lets the next morning’s block start sharp.
The limits of deep-work advice
Here is what most deep-work writing skips: the advice has real edges, and pretending otherwise sets people up to fail. Deep work is genuinely powerful for output that is individual, creative, and hard — and it is a poor fit for a surprising amount of actual work.
Some jobs are paid to be interruptible. A manager, a customer-support lead, an ER physician, a founder in a crisis week — their value is partly availability. Telling a support manager to go monastic for four hours is telling them to stop doing their job. Newport himself concedes this: the monastic and bimodal philosophies are explicitly for a minority of roles, and he caps the recommendation for everyone else at an hour or two a day. The mistake is treating “protect deep work” as a universal maximization target rather than a portfolio decision.
Interruptions are not purely a tax. The deep-work frame treats every interruption as pure loss, and the switching research above backs that up for individual focused tasks. But the fuller picture is more mixed. The most comprehensive review of the field — Puranik, Koopman, and Vough’s 2020 Pardon the Interruption in the Journal of Management, which integrated 247 studies — concludes that interruptions carry documented costs and documented benefits, and that the outcome depends on timing, type, and relevance. Their own work found that being interrupted can increase a sense of belonging and connection at work; other studies find that a well-timed interruption surfaces information the focused worker would have missed. In genuinely collaborative work, some of what looks like fragmentation is the coordination that makes the collective output possible. The lesson is not “interruptions are fine” — it’s that the goal is fewer low-value, badly-timed interruptions, not zero contact with other humans.
The manager’s dilemma. Push this one step further and you get a real tension inside teams. If everyone maximizes their own deep-work blocks and goes dark, the person whose deep work depends on a quick answer from a colleague now waits half a day. Individual focus optimization can degrade the team’s throughput. This is why deep work is as much a scheduling-norms problem as a personal-discipline one — solved better by shared conventions like no-meeting days and agreed response-time expectations than by everyone privately hoarding focus.
Newport’s own second thoughts. It’s telling that Newport’s later thinking moved past pure deep work. In his 2024 book Slow Productivity he names the real enemy as pseudo-productivity — using visible busyness as a proxy for real output — and prescribes not just more focus but doing fewer things and working at a natural pace. The nuance matters: cramming your calendar full of deep-work blocks can become its own version of the overwork trap. Depth without restraint is still a treadmill; it’s just a more concentrated one.
None of this dissolves the core insight. Fragmented attention really does wreck demanding work, and protecting focus really is high-leverage. But the honest version of the advice is conditional: know which of your work is deep, cap the shallow that fragments it, and don’t apply the monastic ideal to a job that is fundamentally about being reachable.
How to measure your own deep work
The way to keep deep work honest — powerful where it belongs, not fetishized — is to measure it, and to measure the right thing. Time at your desk is meaningless; the multiplication formula says intensity is what counts, so the metric that predicts output is hours of genuine, distraction-free focus per week. Newport recommends tracking exactly this number and treating it as the score to improve.
A workable way to run it:
- Count deep-work hours, not clock hours. At the end of each day, log only the blocks where you worked on one demanding task with no switching. Most people are startled by how small the honest number is — often two or three hours on a “busy” day.
- Start low and build. Newport suggests beginners cap total deep work near an hour a day and grow the capacity over weeks. Concentration is a skill; treating it as unlimited from day one guarantees a flameout.
- Design distraction out, not with willpower. Close the email tab, put the phone in another room, silence chat for the block. Leroy’s point is that resisting a ping still costs residue — the only reliable fix is to make the interruption impossible.
- Batch the shallow into named windows. Group email, messages, and admin into one or two defined windows instead of letting them trickle across the whole day. This caps shallowness without pretending you can delete it.
- Watch the trend, not the day. A single bad day tells you nothing. A weekly deep-work-hours line trending up (or collapsing during a heavy-meeting stretch) tells you whether your defenses are actually holding. Our time-blocking guide covers the mechanics of protecting those blocks in the first place.
Track this for a month and you get something rarer than motivation: an honest picture of how much of your paid time is actually spent on the work only you can do — and a number you can defend when someone tries to book over it.
Where Carly fits
The reason deep work fails in practice usually isn’t weak discipline — it’s that the shallow layer keeps fragmenting the day faster than any willpower can hold it back. Email, scheduling back-and-forth, and pings arrive continuously, and each one drags attention residue across your focus blocks. Carly is an AI executive assistant that works across your email, calendar, tasks, and CRM with 200+ integrations, and it acts on triggers — so it can protect deep-work blocks on your calendar, triage and draft your email, and handle scheduling logistics before they interrupt you, keeping the shallow layer from eating your focus. It doesn’t do the concentrating for you; the deep work is still yours. It just clears the fragmentation that was standing in its way. Carly starts at $35/month.
FAQ
What is deep work in simple terms? Deep work is focusing without distraction on a mentally demanding task. Cal Newport defines it as professional activity performed in a state of distraction-free concentration that pushes your cognitive capabilities to their limit — effort that creates new value and is hard to replicate. Its opposite is shallow work: routine, logistical tasks done while distracted.
Why does switching tasks cost so much? Because attention has inertia. Sophie Leroy’s research on attention residue shows part of your mind stays stuck on the previous task after every switch, and Rubinstein, Meyer, and Evans found that the switch itself carries a time cost that grows with complexity. Each individual toggle feels free, but the APA notes the aggregate drag can reach 40 percent of productive time.
Is it really 23 minutes to refocus after an interruption? Treat that number with caution. The “23 minutes and 15 seconds” figure comes from a 2006 Gallup interview with researcher Gloria Mark, not from a peer-reviewed paper. Mark’s published work does show that refocusing is slow and that the day is heavily fragmented, but the precise stopwatch figure is an illustration, not a measured law.
Are interruptions always bad for work? No — and this is where a lot of deep-work advice overreaches. For individual focused tasks, interruptions are costly. But the largest review of the research, Puranik, Koopman, and Vough’s 2020 Pardon the Interruption, found interruptions carry both costs and benefits depending on timing, type, and relevance, including a greater sense of connection at work. The goal is fewer low-value interruptions, not zero contact.
Who shouldn’t try to work monastically? Anyone whose value depends on being reachable — managers, customer support, clinicians, founders mid-crisis. Newport limits the monastic and bimodal philosophies to a minority of roles and caps deep work at an hour or two a day for most people. Deep work is a portfolio decision, not a universal maximization target.
How should I measure my deep work? Track hours of genuine, distraction-free focus per week — not time at your desk. Because output equals time multiplied by intensity, focused hours are the number that actually predicts results. Log only the blocks with no task-switching, start near an hour a day, and watch the weekly trend rather than any single day.
Related: best AI tools for deep work, time-blocking guide, context switching statistics, no-meeting days guide, executive time management secrets
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
