Cost Of
Cost of context switching
Each interruption costs a recovery. Multiply by people and by working days.
What context switching costs a team
An interruption costs far more than the interruption. Gloria Mark's field studies at UC Irvine found people took about 23 minutes on average to get back to interrupted work, usually after handling two other things first. Six interruptions in a day is a quiet day for most people — and it is more than two hours of recovery per person, every day.
This calculator turns that into money: interruptions per person per day × recovery minutes × people × working days, priced at fully-loaded hourly cost. Recovery time is capped at one working day per person, so the total stays defensible even at extreme inputs.
How the number is built
- Per person per day
- interruptions × recovery minutes, capped at 480 minutes — one working day is the maximum anyone can lose.
- Per year
- × people × working days (240 by default: 48 weeks × 5 days).
- In money
- hours lost × fully-loaded hourly cost (salary × multiplier ÷ working hours per year).
- Also shown
- the share of everyone's working time this consumes, the full-time-equivalent headcount it matches, and what halving interruptions would keep.
When to use it
-
Proposing no-meeting blocks
A focus-time policy is easier to approve when the status quo has a price. Compare the current interruption rate against a realistic target and show the difference.
-
Reviewing alerting and on-call
Pages that could have been a ticket are interruptions with a salary cost. Price the current volume before deciding how many alerts to keep.
-
Sizing the cost of a reorg or a tool change
More handoffs mean more interruptions. Estimating the new rate gives you the ongoing cost of a structure, not just the cost of the transition.
Questions
Is 23 minutes of recovery realistic?
It is the average from Mark, Gonzalez & Harris (CHI 2005) and the associated interviews, and it covers returning to the original task after handling other work in between. Deep, focused work tends to be worse; routine work recovers faster. The field is editable — try 10 minutes and see whether the argument still holds.
Why is the loss capped at one workday per person?
Because recovery cannot exceed the hours in the day. Without a cap, a high interruption count multiplied by a long recovery produces a number larger than the working day itself, which is easy to attack and easy to dismiss.
Does an interruption cost the recovery time twice?
No. Only the recovery minutes are counted, not the interruption itself — a two-minute question with a 23-minute recovery is priced at 23 minutes. That deliberately understates the total.
How many interruptions do people actually get?
It varies enormously by role. Six a day is a common baseline for engineering-style work; support, sales and management roles are routinely in the double digits. Count a typical day for the people affected rather than the worst one.
Can I use this for meetings too?
Use the meeting calculator for the meeting itself, and its switching-overhead field for the recovery afterwards. This calculator is for the unscheduled interruptions in between.