In 2009, Sophie Leroy at the University of Minnesota published a paper that should have restructured every calendar in every knowledge-work organisation on the planet. The finding: when you switch from Task A to Task B, part of your cognitive processing remains stuck on Task A. Leroy called this "attention residue." The residue is not a feeling. It is a measurable degradation in performance. Subjects who switched tasks while Task A was incomplete performed significantly worse on Task B than subjects who completed Task A before switching. The attention didn't transfer cleanly. It leaked.
The residue persists for 15 to 25 minutes. During that window, you are cognitively impaired — not dramatically, not in a way you'd notice subjectively, but in a way that shows up in error rates, decision quality, and creative output. You feel like you're working on Task B. You are actually running two cognitive processes: working on Task B while your mind continues processing the unresolved elements of Task A. The result is neither task gets your full capacity. You are half-present twice rather than fully present once.
Cal Newport built his entire Deep Work thesis on this foundation. His argument: knowledge workers who fragment their attention across tasks, emails, Slack messages, and meetings never reach the state of deep cognitive engagement where the highest-value intellectual work happens. They can't — the attention residue from each interruption degrades the next 15-25 minutes of cognitive function. By the time the residue clears, the next interruption arrives. The worker spends the entire day in a degraded cognitive state, producing shallow output at a fraction of their potential.
Gloria Mark at UC Irvine quantified the fragmentation. Her research found that the average knowledge worker switches tasks every 11 minutes. Not every hour. Not every 30 minutes. Every 11 minutes. And after each switch, Leroy's research predicts 15-25 minutes of residue. The math is devastating: if you switch every 11 minutes and the residue lasts 15 minutes, you never clear the residue from the previous switch before the next one arrives. You spend the entire workday in a state of continuous partial attention — never fully engaged with anything, always carrying cognitive debris from the last thing you touched.
Microsoft Research confirmed the recovery cost from a different angle. Their study found that after a single email interruption, workers took an average of 23 minutes to return to the original task — not to reach peak performance on the task, but simply to re-engage with it at all. Many workers never returned. They got pulled into a cascade of related tasks triggered by the interruption and ended the day having abandoned the original work entirely.
The strategic implication is stark. Time-blocking, maker schedules, meeting-free days, and notification discipline are not productivity hacks. They are not personal preferences or lifestyle choices. They are structural defences against a cognitive vulnerability that degrades every knowledge worker's output every day. An organisation that fragments its workers' attention with constant interruptions is not just annoying its employees. It is systematically preventing them from doing their best work.
Section 2
How to See It
Attention residue is invisible in the moment and obvious in the outcome. You see it in the gap between how busy someone is and how much they produce, in the difference between hours worked and value created, and in the chronic inability of talented people to produce work that matches their capability.
You're seeing attention residue when high-performing individuals produce mediocre output, when teams are constantly busy but perpetually behind, or when the quality of decisions degrades as the day progresses without any change in the difficulty of the decisions.
Software Engineering
You're seeing attention residue when a senior engineer who can architect complex systems in a weekend hackathon produces middling code during the workweek. The difference is not effort or motivation. The weekend has six-hour uninterrupted blocks. The workweek has standups at 9:15, a Slack culture that expects sub-five-minute response times, "quick sync" meetings scattered across the afternoon, and a code review queue that pings every 30 minutes. The engineer never enters a deep focus state during business hours. The residue from each interruption stacks on top of the last, producing a full workday of degraded cognitive performance.
Executive Leadership
You're seeing attention residue when a CEO makes sharp strategic decisions at a quarterly offsite but makes reactive, inconsistent decisions during normal weeks. The offsite provides three days of single-topic immersion — no emails, no operational fires, no context-switching. The normal week provides 30-minute windows between back-to-back meetings, each on a different topic. The CEO carries residue from the pricing discussion into the hiring review, residue from the hiring review into the product roadmap session, and residue from the product roadmap into the investor call. Each decision is contaminated by the cognitive leftovers of the previous one.
Content & Media
You're seeing attention residue when a writer produces a brilliant 3,000-word essay in a single four-hour morning session but cannot produce anything usable across five fragmented one-hour blocks totalling the same time. The total hours are identical. The output is radically different. Writing requires sustained attention to hold the entire structure of an argument in working memory while simultaneously generating and evaluating prose. Attention residue from a midday meeting dumps the argument structure from working memory. The writer spends 20 minutes reloading context after each interruption, leaving 40 minutes of degraded production per block.
Product Management
You're seeing attention residue when a product manager toggles between customer calls, sprint planning, stakeholder updates, and roadmap reviews across a single afternoon and ends the day unable to articulate a clear product direction. Each context requires a different mental model — the customer's pain, the engineering team's capacity, the stakeholder's priorities, the market's trajectory. Switching between them every 30 minutes means the PM never fully inhabits any perspective. The residue from the customer call bleeds into sprint planning. Engineering constraints from the sprint bleed into the stakeholder update. The PM becomes a router of information rather than an integrator of insight.
Section 3
How to Use It
Attention residue is a constraint to be managed, not a flaw to be fixed. You cannot eliminate the residue — it is how human cognition works. You can structure your time, your environment, and your organisation to minimise the frequency of task-switching and maximise the length of uninterrupted blocks.
Decision filter
"Before scheduling any meeting, interruption, or task-switch: is the value of this interruption greater than the 15-25 minutes of degraded performance it will impose on the person being interrupted? If no, batch it. If yes, schedule it at a natural transition point — between deep-work blocks, not in the middle of one."
As a founder
Design your day around two or three deep-work blocks of 90-120 minutes each. Protect them with the same ferocity you'd protect a meeting with your most important investor. No Slack. No email. No "quick questions." Move all communication to designated batching windows — 30 minutes at 11am, 30 minutes at 3pm. The resistance will be cultural: your team will feel you're less available. The output will be structural: the quality of your thinking will improve measurably within two weeks. Paul Graham's "Maker's Schedule, Manager's Schedule" essay is the operating manual — makers need half-day blocks, and a single meeting in the middle of a maker's afternoon destroys the entire afternoon.
As a team lead
Audit your team's calendars for fragmentation. Count the number of times each person switches contexts in a day — meetings, Slack channels, code reviews, standups, one-on-ones. If the average team member switches more than six times per day, your team is operating in chronic residue. Consolidate meetings into designated windows. Institute "focus hours" where Slack expectations shift from "respond in 5 minutes" to "respond by end of day." Track the impact on output quality, not output quantity — attention residue doesn't reduce the volume of work; it reduces the calibre.
As a decision-maker
Structure your decision calendar to batch similar decisions together. Moving from a hiring decision to a pricing decision to a product-kill decision in 90 minutes guarantees that each decision carries residue from the last. Instead, batch all hiring decisions into a single session. Batch all product decisions into another. The cognitive context loads once and serves multiple decisions rather than loading and unloading repeatedly. Jeff Bezos reportedly makes all high-stakes decisions before lunch, when cognitive resources are freshest and residue from the day's accumulated switches hasn't yet degraded his processing.
Common misapplication: Treating attention residue as an excuse for unresponsiveness. Protecting deep-work blocks does not mean becoming unreachable. It means being deliberately unreachable during specific windows and deliberately responsive during others. The discipline is time-boxing both modes — not choosing one permanently.
Second misapplication: Assuming longer blocks are always better. Research on ultradian rhythms suggests that humans cycle through periods of high and low alertness in roughly 90-minute intervals. A four-hour uninterrupted block may hit diminishing returns after 90-120 minutes as natural cognitive fatigue sets in. The goal is not maximum block length but minimum switching within the block.
Section 4
The Mechanism
Section 5
Founders & Leaders in Action
The leaders below recognised that attention is not a renewable resource that replenishes between meetings. It is a finite, degradable asset that must be structurally protected — not through willpower but through calendar architecture and organisational design.
Lütke built Shopify's engineering culture around a principle that directly addresses attention residue: meetings are a last resort, not a default. Shopify eliminated recurring meetings that lacked clear decision-making purpose, shifted to asynchronous communication as the primary mode, and gave engineers the autonomy to structure their days around deep-work blocks. Lütke's philosophy: "The maker's schedule is not a preference. It is a prerequisite for the kind of work we need our engineers to do." Shopify's internal tools reflect this — async updates replace standups, written memos replace brainstorming meetings, and engineers can decline any meeting that doesn't have a clear agenda and decision to be made. Lütke himself models the behaviour: his calendar is sparse by CEO standards, with large blocks reserved for coding, thinking, and product review. The organisational result is a company that ships at a pace inconsistent with its size — because its builders spend most of their time building rather than switching between meetings about building.
Bezos engineered Amazon's meeting culture to minimise attention residue at the executive level. His "two-pizza team" rule — no team should be larger than two pizzas can feed — reduced the number of coordination meetings by reducing the number of people who needed to coordinate. His six-page memo requirement replaced PowerPoint presentations with written narratives that must be read silently at the start of each meeting. The silent reading period serves a specific attention-residue function: it forces every participant to load the same cognitive context before discussion begins, eliminating the residue that participants carry from their previous meetings. A VP who rushes from a supply chain review into a product strategy meeting without the memo arrives carrying residue from the supply chain discussion. The memo's silent reading period clears that residue and replaces it with the current meeting's context. Bezos also famously refused to schedule important meetings before 10am and limited his high-IQ decision meetings to before lunch — structuring his day so that the decisions requiring the most cognitive clarity receive the least residue-contaminated attention.
Section 6
Visual Explanation
The left panel shows a typical fragmented workday: five context switches before mid-morning, each generating 15-25 minutes of residue that never clears before the next switch arrives. Cognitive capacity hovers around 35%. The right panel shows the same eight hours structured around a protected deep-work block: one switch at a natural transition point, residue clears fully, and cognitive capacity reaches 85%. Same person. Same hours. Radically different output. The equation at the bottom makes the arithmetic explicit: switching every 11 minutes plus residue lasting 15-25 minutes equals a permanent state of cognitive impairment. The only defence is structural — blocking time, batching communication, and protecting the transitions between modes of work.
Section 7
Connected Models
Attention residue sits at the intersection of cognitive science, organisational design, and productivity strategy. The connected models explain why deep focus is valuable, why organisations fragment it, and what structural defences exist against the fragmentation.
Reinforces
Deep Work
Cal Newport's Deep Work is the strategic framework built on attention residue's cognitive foundation. Newport's argument: the ability to perform deep work — sustained, undistracted focus on cognitively demanding tasks — is becoming both more rare and more valuable in the knowledge economy. Attention residue explains why it's rare: every interruption generates residue that degrades the next 15-25 minutes, and modern work environments generate interruptions faster than residue clears. Deep work is not a personality trait. It is a structural achievement that requires defending against the forces that generate residue.
Reinforces
[Flow State](/mental-models/flow-state)
Mihaly Csikszentmihalyi's flow state requires 15-25 minutes of uninterrupted engagement to enter. Attention residue takes 15-25 minutes to clear. The alignment is not coincidental — flow requires the full allocation of working memory to a single task, and residue occupies working memory slots that prevent full allocation. A single interruption doesn't just break flow; it prevents re-entry for the duration of the residue. In a fragmented workday, flow becomes structurally impossible. The worker never clears the residue long enough to enter the state where their highest-quality work happens.
Tension
[Context](/mental-models/context) Switching
Context switching is the action. Attention residue is the cost. Most discussions of multitasking focus on the switching itself — the time lost in transition. Attention residue reveals that the cost extends far beyond the transition: you pay the switching cost during the transition and the residue cost for 15-25 minutes after. A "quick" two-minute Slack exchange costs two minutes of switching plus 15-25 minutes of residue. The true cost of the interruption is 10-12x the visible duration.
Section 8
One Key Quote
"If you service every interruption as it arrives, you'll never produce anything of real value. The ability to perform deep work is becoming increasingly rare at exactly the same time it is becoming increasingly valuable in our economy."
— Cal Newport, Deep Work: Rules for Focused Success in a Distracted World (2016)
Newport's argument is that attention residue creates a market inefficiency. Most knowledge workers operate in chronically fragmented environments, producing output far below their potential. The few who structure their work to minimise residue — who protect deep-work blocks, batch communication, and refuse to service every interruption as it arrives — produce disproportionately higher-quality work. The competitive advantage is not intelligence or effort. It is calendar architecture. The worker who blocks four hours of uninterrupted focus each morning will outperform the equally talented worker who checks email every 15 minutes — not because they work harder, but because they work in a cognitively uncorrupted state.
The economic implication: organisations that protect their knowledge workers' attention will systematically outperform organisations that fragment it. This is not a marginal effect. Leroy's research suggests that the performance gap between a residue-free state and a chronic-residue state is 30-40% on cognitively demanding tasks. An organisation that reduces attention fragmentation by half is not incrementally improving productivity. It is unlocking a third of its workforce's latent capacity.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Attention residue is the single most underpriced concept in knowledge work. Every organisation talks about productivity. Almost none of them address the cognitive mechanism that destroys it. They invest in better tools, better processes, better talent — and then fragment that talent's attention with meeting cultures, instant-messaging norms, and open-plan offices that guarantee the talent never operates at capacity. It is like buying a Formula 1 car and driving it in first gear because you never gave it enough runway to accelerate.
The math is not subtle. Gloria Mark's data: switches every 11 minutes. Leroy's data: residue lasts 15-25 minutes. Microsoft's data: 23 minutes to re-engage after a single email interruption. These are not contested findings. The research is replicated and robust. The average knowledge worker never reaches deep focus during a typical workday. They spend eight hours in a degraded cognitive state and produce output that reflects it. Then they go home and do their real thinking at night or on weekends — in the uninterrupted hours that their employer's meeting culture denied them during business hours.
Shopify understood this.Tobi Lütke's decision to eliminate unnecessary meetings and default to async communication was not a cultural preference. It was a structural response to attention residue. When Shopify engineers can protect four-hour coding blocks, they produce architecture. When they're interrupted every 15 minutes, they produce patches. The same engineers. The same skills. Radically different output based entirely on whether their attention was protected or fragmented.
The meeting-industrial complex is the primary residue generator in most organisations. The average manager spends 23 hours per week in meetings (Harvard Business Review, 2022). Each meeting generates residue that contaminates the time between meetings. A manager with six 30-minute meetings spread across a day has six windows of residue-contaminated time between meetings — windows too short to clear residue and too fragmented for deep work. The manager is busy for eight hours and effective for perhaps two. The meetings create the illusion of productivity while destroying the conditions for it.
Remote work didn't fix the problem. It relocated it. The open-plan office's interruptions were replaced by Slack's notification stream. The tap-on-the-shoulder was replaced by the @mention. The ambient noise was replaced by video-call fatigue. The medium changed. The fragmentation didn't. Many remote workers report more interruptions than they experienced in the office — because digital communication tools have zero friction. Sending a Slack message costs the sender nothing. It costs the recipient 15-25 minutes of residue.
Section 10
Test Yourself
The scenarios below test whether you can identify attention residue in organisational contexts, quantify its cost, and design structural defences against it.
Each scenario presents a common workplace situation — your job is to determine whether attention residue is degrading performance and what structural change would address it.
Spot the residue.
Scenario 1
A software engineering team has a daily standup at 9:15am (15 minutes), a team sync at 11:00am (30 minutes), sprint planning at 2:00pm (1 hour), and ad-hoc Slack messages throughout the day with an expected response time under 10 minutes. The tech lead complains that the team ships fewer features per sprint than a smaller team at a competing company that has half the engineers.
Scenario 2
A CEO schedules her day with 30-minute meetings back-to-back from 9am to 5pm, each on a different topic: board prep, product review, HR issue, sales pipeline, legal review, investor relations, marketing strategy, engineering roadmap. She reports feeling 'mentally exhausted but unproductive' and notes that her strategic thinking happens on weekend mornings when she's alone.
Section 11
Top Resources
Attention residue sits at the intersection of cognitive psychology, organisational behaviour, and productivity strategy. The strongest resources provide the empirical foundation, the strategic framework, and the practical tactics for defending against it.
The origin paper. Leroy's experiments at the University of Minnesota demonstrated that attention residue is real, measurable, and performance-degrading. The paper's most important finding: residue is worse when the previous task was unfinished or time-pressured — which describes virtually every task in a knowledge worker's day. This is the empirical foundation that everything else builds on.
Newport translated Leroy's cognitive research into a strategic framework for knowledge workers. The book's core argument — that the ability to perform deep work is both increasingly rare and increasingly valuable — is built directly on attention residue's mechanism. The practical rules (work deeply, embrace boredom, quit social media, drain the shallows) are structural defences against residue generation. The most widely read treatment of the concept and the most actionable.
Mark's decades of research at UC Irvine quantified workplace fragmentation: the average knowledge worker's attention span on a single screen shrank from 2.5 minutes in 2004 to 47 seconds in 2020. Her finding that workers switch tasks every 11 minutes became one of the most cited statistics in productivity research. This book synthesises 20 years of empirical data on how digital environments fragment attention and what individuals and organisations can do about it.
Graham's essay provides the organisational design framework that attention residue demands. His distinction between makers (who need half-day blocks) and managers (who work in one-hour slots) explains why the same meeting culture that works for managers destroys makers' productivity. The essay is short, widely referenced, and operationally precise — it gives makers the language to defend their deep-work blocks and gives managers the framework to understand why they should.
This study quantified the full cost of workplace interruptions: after being interrupted, workers took an average of 23 minutes and 15 seconds to return to the original task. Many never returned at all, instead getting pulled into cascading tasks triggered by the interruption. The paper provides the empirical basis for understanding why a "quick question" is never quick — the visible interruption is seconds; the cognitive recovery is half an hour.
Attention Residue Timeline — when you switch tasks, cognitive residue from the previous task degrades performance on the new task for 15-25 minutes. Frequent switching means you never clear the residue, operating in a permanent state of degraded cognition.
Tension
Opportunity [Cost](/mental-models/cost)
Every minute spent in attention residue is a minute of degraded cognitive capacity that could have been spent at full capacity. The opportunity cost of fragmentation is not the sum of the interruptions — it is the difference between what was produced in a residue-contaminated state and what could have been produced in a clear state. For knowledge workers whose output quality depends on cognitive clarity, this gap represents the single largest source of wasted potential in their professional lives.
Leads-to
Bottleneck
In knowledge-work organisations, the bottleneck is rarely talent, tools, or information. It is attention. The scarcest resource in a software company is not engineering hours — it is uninterrupted engineering hours. Attention residue explains why adding more engineers to a fragmented environment produces diminishing returns: each new engineer faces the same residue-generating interruption structure. The bottleneck is not people. It is the cognitive environment those people operate in.
Leads-to
Maker vs Manager Schedule
Paul Graham's distinction between maker and manager schedules is a direct organisational response to attention residue. Managers operate in one-hour blocks — their work is meetings, decisions, and coordination, all of which tolerate and even benefit from variety. Makers — engineers, designers, writers — need half-day blocks because their work requires loading complex mental models into working memory and keeping them there. A single meeting in the middle of a maker's afternoon generates residue that contaminates the entire remaining block. Graham's framework doesn't just describe a preference. It describes a cognitive requirement.
The fix is not complicated. It is culturally difficult. Block 2-3 hours of deep work daily. Batch communication into designated windows. Reduce meeting frequency by 50% and replace the eliminated meetings with async written updates. Make the default "no meeting" and require justification for exceptions. These are simple structural changes. They are culturally difficult because they require managers to tolerate delayed responses, leaders to trust async communication, and organisations to value output quality over input visibility. The organisation that makes these changes will outperform. The organisations that don't will continue paying full-time salaries for part-time cognitive engagement.
Scenario 3
A remote product team implements 'Focus Fridays' — no meetings allowed on Fridays. After three months, the team reports that Focus Fridays feel productive but that Monday through Thursday have become worse: more meetings are crammed into four days, Slack volume increased because people batch questions for the communication windows, and several team members report working late on Thursday to 'prepare for Focus Friday.' Overall output has not measurably improved.
