Interruption Cost Calculator

What this tool does

CM8-343 measures what fragmentation costs. For each team and each source of interruption it counts how often it happens, how long it takes, and how long it takes to get back to what was being done — then reports the hours and money lost, the share of the working day consumed, and the average uninterrupted block anybody actually gets.

Everything runs inside this single file: no account, no upload, no network request, which matters because the output is a candid picture of how a team's day really works.

The recovery tail

The insight this tool exists to make visible is that an interruption costs far more than it takes. A three-minute question during difficult work is not three minutes: it is three minutes plus however long it takes to rebuild the state of mind that was interrupted.

The tool charges that recovery separately and shows it as its own block, which is what makes the chart persuasive. Twenty-two alerts a day at one minute each looks like twenty-two minutes and is really an hour and three quarters once four minutes of recovery lands on each one — the recovery being five times the interruption. Nobody would schedule an hour and three quarters of daily meetings without noticing.

Why depth of work matters

Recovery is not a constant, and treating it as one is the commonest way this calculation is done badly.

Interrupting routine work — answering a straightforward query, processing a form, checking a delivery — costs close to nothing beyond the interruption itself. Interrupting analysis, writing, design or diagnosis costs a great deal more, because what is lost is not the place in a task but the whole assembled context.

That is why each row records what is being interrupted, and why the sample shows a customer service team absorbing twenty-eight interruptions a day cheaply while six design engineers lose far more to fourteen. The same volume is not the same cost.

The longest uninterrupted block

Total hours lost is the number people quote. The average uninterrupted block is the number that predicts whether anything difficult will get done.

If a day is broken by twenty interruptions, the remaining time arrives in about twenty-one pieces. Even a team losing "only" a fifth of its day can be left with fifteen-minute fragments — and demanding work does not fit into fifteen minutes at all. Two teams with identical hours lost can be in completely different situations depending on whether those hours arrive as one block or forty.

This is why reducing the number of interruptions usually matters more than reducing their duration, and why moving them rather than removing them works so well.

Where the model breaks down

Recovery time is added once per interruption, and that is only sound while interruptions are far enough apart to recover between them. Above roughly ten to twelve interruptions a day on deep work, this tool overstates the cost — and it is worth knowing exactly why.

If you are interrupted again before you have got back into the task, you do not pay the recovery twice. You never returned to full depth in the first place, so the second interruption costs less than the first. Beyond a certain frequency the cost stops rising and simply saturates: the person is no longer doing demanding work at all, they are doing shallow work continuously.

The tool does not model that saturation, deliberately — doing so would require an assumption about the shape of the curve that nobody can defend, and it would hide the finding behind a formula. Instead, read a team showing more than about half its day lost to deep-work interruptions as saying something simpler and more useful: this team is not doing deep work, whatever its job description says. That conclusion is correct, and it does not depend on the arithmetic being precise.

The same caution applies in reverse for roles that are interruption-driven by design. Customer service and supervision score high volumes on shallow work with near-zero recovery, which is exactly right — their cost per interruption genuinely is low, and the total is the job rather than a problem.

Measuring rather than guessing

People underestimate how often they are interrupted, consistently and by a wide margin. A tally sheet for a single representative day is almost always a surprise, and it takes about thirty seconds of anybody's attention to keep.

Mark rows as measured or estimated so the report is honest about which is which. An estimated row is fine for a first pass; it is not a basis for changing how a team works, and the sample data deliberately includes both.

What is actually avoidable

Avoidable does not mean the work disappears. It means the interruption could be batched, answered by information that already exists, routed to somebody else, or prevented at source.

Be honest with this figure — an inflated one produces a saving nobody can deliver, and the exercise loses credibility on the first attempt. Thirty to fifty per cent is realistic for most discretionary sources. The tool refuses 100% outright: if something genuinely serves no purpose at all, that is not an interruption problem, it is something to stop.

The most avoidable source is usually the one caused by a recurring failure. Escalations, chasing and "quick questions" about the same three things are not interruptions to be managed; they are symptoms, and fixing the cause removes them permanently.

Batching beats banning

The change that works is almost never "fewer interruptions". It is the same volume, in predictable windows.

Two open windows a day instead of open access serves the same questions, keeps the team helpful, and converts a fragmented day into two long blocks. Nobody is refused; they simply wait until eleven. The same logic applies to messages answered at set times, alerts summarised hourly rather than fired instantly, and standing meetings collapsed into one block on one day.

Record the intended change in the "what would reduce it" column so the report carries an action rather than a complaint.

What not to conclude

Interruptions are how organisations coordinate. A team that cannot be interrupted is a team that has stopped helping anybody, and the cost of that lands somewhere else — usually on people waiting for an answer they could have had in thirty seconds.

Some roles are interruption-driven by design. The sample includes production supervisors, whose whole job is being available, and the row is there to make the point: measuring the cost of something is not the same as proposing to remove it. Knowing what availability costs is useful when deciding how many people need to be available, and to whom.

And, as ever, hours released are not cash. Nobody is made redundant because interruptions fell. What is released is the ability to do work that needs concentration — valuable, and only if the released time is protected rather than immediately refilled.

The formulas

The average block treats interruptions as evenly spread through the day, which they are not — they cluster around the start of the morning and after meetings. Real blocks are therefore both longer and shorter than the average suggests, and a team whose average looks workable can still have no usable block on a bad day.

FAQ

What recovery time should I use? Observe it. Ask two people to note, for one day, how long it takes to get properly back into something after being interrupted. The defaults here are moderate assumptions, and published research figures vary enormously because they were measured on very different work.

Should meetings count as interruptions? Yes, when they break into work that was already under way — which is most of them. A meeting at nine o'clock interrupts nothing; the same meeting at half past two costs the morning's second half as well as its own hour.

The total looks implausibly large. Check the times a day first, then the recovery setting. Multiplying a slightly generous recovery figure by a slightly generous frequency across a whole team produces a number that will be challenged, and it should be.

Can I use this to justify home working, or an office move? It is a legitimate input, but a weak argument on its own — remote work substitutes some interruption sources for others rather than removing them. Measure both sides before claiming a difference.

One team's row says the whole day is gone. The tool refuses to record more than a full day from one source. If that really is the position, the finding belongs in a conversation about what that team is for, not in a spreadsheet.

Saving your work

Sources, settings and the report header are written to this browser's local storage as you type. Treat Export .json as the real save, which Import .json restores anywhere. Export CSV gives you the measurements for spreadsheet work. Reset asks twice, then erases everything. There is no undo.

Accuracy & disclaimer

The number carrying this whole calculation — recovery time — is not a constant, and published figures for it vary enormously because they were measured on different work. Use your own observation and set it deliberately.

Be careful with the conclusion, too. Interruptions are how organisations coordinate, and the finding worth acting on is almost never "fewer interruptions" — it is that the same volume, batched into predictable windows, costs a fraction of what it costs scattered through the day.