Context Switching: What a Blocker App Can Fix, and What It Structurally Can't

Which one you need depends on who is doing the interrupting. If a notification, a tab or a colleague pulls you away, a focus app fixes that today for a few dollars a month. If you close the document yourself because you would rather not be in it, no app can see that, and a coaching conversation is the tool that can. The two are not grades of the same product; they act on different halves of the problem. In the study most often cited on this topic, about half the interruptions the researchers logged came from outside the person and about half came from the person. A blocker acts on the external half because that is the half it can see: a domain name, an app, a clock.

IX Coach works on the other half. It is built as one continuing conversation that carries what you said last session into the next one, so a pattern you noticed on Tuesday is still on the table on Friday instead of resetting. A blocker keeps no account of your reasons because reasons are not among its inputs. That is a structural difference between the two categories, not a verdict on which you need; the one-week test further down shows which half is yours.

What the field research actually measured

The study most often cited on this subject is Gloria Mark, Victor González and Justin Harris, No Task Left Behind? Examining the Nature of Fragmented Work, published at ACM CHI 2005. It is an observational field study: 24 information workers at a single outsourcing firm, shadowed over a thirteen-month period, with more than 700 formal hours of observation recorded to the second.

The numbers it reports are these. Informants spent an average of 11 minutes 4 seconds in a working sphere — the researchers' term for a coherent unit of work, like preparing a proposal — before switching or being interrupted. 57.1% of those segments were interrupted. 77.2% of interrupted work was picked back up the same day, and when it was, it took an average of 25 minutes 26 seconds to get back to it. The standard deviation on that last figure is 54 minutes 48 seconds, which is larger than the average and is the part almost never quoted: the spread is enormous, and a single average conceals it. Before resuming, informants had passed through an average of 2.26 other working spheres — the return trip is rarely direct.

One more finding from the same paper cuts against the usual story. The researchers noted that people who resumed a task on their own initiative had worked significantly longer in it before being interrupted, and connected this to the Zeigarnik effect — the tendency for unfinished work to stay active in mind and pull you back. Interruption and resumption are not only a cost to be minimised; some of what pulls you back is the same mechanism that makes an unfinished thing hard to drop.

Explore: attention residue

About half the interruptions came from inside

The finding that decides this comparison is in Table 3 of that paper. Interruptions were coded as external — a phone ringing, a colleague at the cubicle, an email arriving — or internal, meaning the person stopped of their own volition. Across all informants the split was 52.0% internal to 48.0% external. It varied by role: managers experienced more external interruptions (59.2%) than internal ones (40.8%), while analysts and developers experienced the two about equally.

Read against a purchase decision, that number is the whole thing. Software can intercept a notification, a site, or an app. Roughly half of what broke these workers' concentration was not a notification, a site, or an app. It was a person deciding, for reasons the study did not have to name, to go and do something else. No blocker has visibility into that decision. The products describe themselves as blocking distractions, which is an accurate description of what they do.

Explore: notification detox · attention economy defense

Two widely quoted numbers that are not in the papers they are credited to

Two figures turn up again and again in articles on this subject, and neither can be found where it is credited.

The first is "23 minutes and 15 seconds to get back on task," attributed almost universally to the 2005 Mark, González and Harris study. It is not in that paper, which reports 25 minutes 26 seconds. The 2008 follow-up by the same lead author contains no task-resumption time at all. Neither paper offers a source for 23:15, so treat it as unsourced.

The second is that task switching can cost up to 40% of productive time, generally credited to Rubinstein, Meyer and Evans, Executive Control of Cognitive Processes in Task Switching, Journal of Experimental Psychology: Human Perception and Performance, 2001. The paper reports something narrower and more careful: task alternation produced switching-time costs that grew with rule complexity and shrank when participants were cued about the upcoming task, measured in laboratory trials of arithmetic and classification. The 40% figure is not among its reported results; it circulates through an association press summary of the work. That does not make the number wrong, only unverified, and an unverified number is not evidence.

What a blocker app does, and what it costs

The category is mature and the products are good at their stated job. Freedom is an internet, app and website blocker that says it "blocks distractions across all your devices at once," so a session started on a laptop holds when you pick up the phone. As of October 8, 2026, its published pricing is a free tier at $0, Premium at $3.33 a month billed yearly or $8.99 billed monthly, and a one-time "Forever" option at $99.50. RescueTime, which pairs blocking with automatic time tracking, prices its Solo "Focus" plan at $7 a month billed annually or $9 billed monthly, with a 14-day trial, and describes the product as helping you "protect your attention and block distractions."

Where these tools genuinely win is the case a coaching conversation cannot reach at all: the reflex. If the interruption is a notification firing, a tab that opens itself out of habit, or a phone that is simply present, then the cheapest correct fix is a piece of software that removes the option, and it works immediately and costs less than a coffee a month. That is a real result and no amount of reflection substitutes for it. Anyone whose interruptions are overwhelmingly external — the managers in the study, at 59.2% — should probably try the app first and may need nothing else.

Explore: notification batching · task batching

What a coaching conversation asks that an app cannot

The internal half is a different kind of problem, and the 2008 experiment is the clearest evidence of why. Mark, Gudith and Klocke interrupted 48 people in a controlled email task and found the opposite of what most people expect: interrupted work was completed faster, not slower — 20.31 and 20.60 minutes under the two interruption conditions against 22.77 minutes uninterrupted — with no measured difference in quality. People compensated by working faster. The cost showed up somewhere else entirely. In the authors' words, subjects experienced "more stress, higher frustration, time pressure and effort," and after only twenty minutes of interrupted performance those measures were significantly higher.

That relocates the problem. If the damage from fragmented work lands mainly as stress, pressure and effort rather than as lost minutes, then a tool that optimises minutes is measuring the wrong thing, and a person who blocks every site and still feels shredded at six o'clock has not failed at using the app. A coaching conversation works on what the blocker has no input for: what the switch is doing for you, what you are avoiding when you leave the document, which of your commitments are actually in conflict, and what you will do differently on Thursday. IX Coach holds that thread across sessions rather than starting from nothing each time, which matters here specifically because self-interruption is a pattern and a pattern is only visible across weeks.

The honest limit: a coach, human or AI, cannot stop your phone from buzzing. If nothing about your reasons needs examining and the problem is purely environmental, coaching is the more expensive way to solve a cheap problem.

Explore: the zeigarnik effect

Choosing between them without buying both

Both categories describe themselves in nearly the same language (focus, attention, deep work), so a feature list will not separate them. One week of observation will, and it costs nothing. Keep a note of every time you leave a task, and write only whether something arrived or you left on your own. That is the same distinction the researchers coded, and you can code it about yourself.

Choose the app if most of your marks are "something arrived." It is the direct fix, it is inexpensive, and it works on the first day. Choose coaching if most of them are "I left" — because in that case a blocker will lock the doors on a room you were leaving for reasons of your own, and you will find another door. Choose both, in that order, if the answer is genuinely mixed: block the external half first, because it is quick and it clears the noise, then look at what is still happening once the notifications are gone. Whatever remains is the internal half.

Choose a human coach or a therapist over either if what is actually going on is that you cannot face the work — where the difficulty is not attention but something the work itself is standing in for. That is a different conversation, and both an app and an AI coach are the wrong shape for it.

Explore: choice overload · technostress

When neither of these is the right call

If difficulty concentrating has been persistent, is affecting your work and your relationships, and has not shifted regardless of what you try, that is a conversation for a licensed clinician rather than for a productivity tool or a coach of any kind — including this one. Neither category is clinical care and neither should be treated as a substitute for it.

And if you are in crisis right now, do not work through a decision like this one. In the United States, the 988 Suicide & Crisis Lifeline is reachable by calling or texting 988. The Crisis Text Line is reachable by texting HOME to 741741. For an immediate emergency, call 911.

Frequently Asked Questions

How long does it really take to get back to a task after an interruption? In the 2005 CHI field study of 24 information workers, interrupted work that was resumed the same day took an average of 25 minutes 26 seconds to return to — but the standard deviation was 54 minutes 48 seconds, larger than the average itself, so treating it as a reliable per-interruption cost overstates what the data supports. The widely-quoted "23 minutes 15 seconds" does not appear in that paper.

Do focus and website blockers actually work? For externally-triggered interruptions, yes — removing the option is a direct fix and a blocker does it immediately. In the 2005 field study, external interruptions were 48.0% of the total, and 59.2% for the managers observed. The limit is the other half: 52.0% of interruptions overall were internal, meaning the person stopped of their own volition, and no blocker has visibility into that.

Does being interrupted make you slower? Not in the controlled experiment built to test it. Mark, Gudith and Klocke (CHI 2008) found 48 subjects completed interrupted tasks faster than uninterrupted ones — 20.31 and 20.60 minutes versus 22.77 — with no difference in quality, because people compensated by working faster. The cost registered as stress, frustration, time pressure and effort instead, and did so within twenty minutes.

What does a context-switching or focus app cost? As of October 8, 2026, from each company's own pricing page: Freedom lists a free tier, Premium at $3.33/month billed yearly or $8.99/month billed monthly, and a one-time $99.50 "Forever" plan. RescueTime lists its Solo Focus plan at $7/month billed annually or $9/month billed monthly, with a 14-day trial.

Can a coach help with attention problems an app can't? With the self-interruption half, which is where an app has no input. A coaching conversation works on why you leave a task rather than on what you leave it for — what the switch relieves, which commitments are in conflict, and what changes next week. Because self-interruption is a pattern rather than an event, it only becomes visible across sessions, which is why continuity between conversations matters more here than in most coaching topics.

Research

• Gloria Mark, Victor M. González, Justin Harris, (2005), No Task Left Behind? Examining the Nature of Fragmented Work, Proceedings of ACM CHI 2005, pp. 321-330 — Primary source. Field observation of 24 information workers at one outsourcing firm over a 13-month period, more than 700 formal hours of observation. States: informants spent an average of 11 min. 4 sec. (sd 18 min. 9 sec.) in a working sphere before switching or being interrupted; 57.1% of working sphere segments were interrupted; 77.2% of interrupted work was resumed the same day; when resumed the same day it took an average of 25 min. 26 sec. (sd 54 min. 48 sec.); before resuming, informants worked in an average of 2.26 other working spheres. Table 3 column average: 52.0% of interruptions internal, 48.0% external. The figure '23 minutes 15 seconds' does not appear in this paper. verified 2026-10-08
• Gloria Mark, Daniela Gudith, Ulrich Klocke, (2008), The Cost of Interrupted Work: More Speed and Stress, Proceedings of ACM CHI 2008 — Primary source. Controlled lab experiment, 48 subjects, simulated email task. Abstract states people 'completed interrupted tasks in less time with no difference in quality' and that they 'compensate for interruptions by working faster, but this comes at a price: experiencing more stress, higher frustration, time pressure and effort.' Table 1 mean time to perform task: 22.77 min uninterrupted, 20.31 min same-context interruption, 20.60 min different-context. Discussion states: 'After only 20 minutes of interrupted performance people reported significantly higher stress, frustration, workload, effort, and pressure.' This paper contains no task-resumption time figure. verified 2026-10-08
• Joshua S. Rubinstein, David E. Meyer, Jeffrey E. Evans, (2001), Executive Control of Cognitive Processes in Task Switching, Journal of Experimental Psychology: Human Perception and Performance, Vol. 27, No. 4, 763-797 — Primary source. Four laboratory experiments in which participants alternated between classification and arithmetic tasks. Abstract states: 'Task alternation yielded switching-time costs that increased with rule complexity but decreased with task cuing.' The widely-quoted '40% of productive time' figure is not in the abstract; it circulates through an APA press summary rather than the paper. verified 2026-10-08
• RescueTime, (2026), Pricing, rescuetime.com — Company's own pricing page. Solo 'Focus' plan listed at $7/month billed annually or $9/month billed monthly; 'Solo+' at $12/month annually or $15/month monthly; 14-day free trial on paid plans. Describes the product as letting you 'protect your attention and block distractions.' verified 2026-10-08
• Freedom, (2026), Pricing, freedom.to — Company's own pricing page. Free tier at $0; Premium listed at $3.33/month billed yearly, $8.99/month billed monthly, or $99.50 one-time for 'Forever.' Describes itself as an 'Internet, App and Website Blocker' and states 'Freedom blocks distractions across all your devices at once.' verified 2026-10-08