The real cost of inbox overload
What email actually takes from a working day, according to research that can be checked, and why the usual statistics quoted about it cannot be.
5 min read
Everyone agrees email is expensive. Almost nobody can tell you what it costs, because the numbers in circulation are mostly unsourced. Search for email statistics and you will find the same handful of figures repeated across dozens of sites, each citing the one before it, none citing anything you can open.
So this article uses a smaller set of numbers. Every one of them comes from a named study you can read yourself, and where a figure is old, it says so.
The number nobody can check
The most quoted statistic about email is that the average professional receives somewhere between 117 and 126 messages a day. It is attributed to industry reports that sit behind a paywall, which means the figure has been repeated for years by people who have not read its source and could not if they tried.
Two others get quoted constantly: that 85% of emails are read in under fifteen seconds, and that workers read four emails for every one they send. Both are commonly attributed to a large 2025 workplace study. They do not appear in it.
None of that means email is cheap. It means the case for it being expensive has been made badly, and a bad case is easy to dismiss.
What can actually be verified
Microsoft's 2025 Work Trend Index is unusual because it combines a survey of 31,000 knowledge workers across 31 markets with aggregated product signals, and publishes its method. Its finding about interruption is specific:
275
Interruptions per day, from meetings, emails and chats. That is roughly one every two minutes across core working hours.
Microsoft, 2025 Work Trend Index (2025).
The same report finds that chats sent outside standard hours are up 15% year over year, at an average of 58 messages per person, and that 60% of meetings are unscheduled. The picture is not one of heavy email specifically. It is one of a working day with no uninterrupted stretches left in it.
How much of the week that takes
The most durable figure on email's share of the workweek comes from McKinsey Global Institute, and it is worth knowing that it is from 2012.
28%
Of the workweek spent managing email, for what the report calls interaction workers.
McKinsey Global Institute, The social economy (2012). Cited with its date because it is old, not current.
The same report separately found that close to 20% of the week goes on hunting for internal information or tracking down the colleague who has it. These are two findings, not slices of one pie, and combining them into a single chart of the workweek would imply a breakdown McKinsey never published.
Thirteen-year-old research is not a description of your Tuesday. It is useful as a floor. Nothing about the intervening years suggests the number went down.
Why interruption costs more than the minutes
The intuitive model is that email costs you the time you spend on it. Research on interrupted work suggests the real cost sits elsewhere.
Gloria Mark and colleagues at UC Irvine studied what happens when work is interrupted, and found that people largely compensate by working faster. The task still gets done. What changes is the experience of doing it: more stress, more frustration, more time pressure, more effort. The output holds and the person absorbs the difference.
A later study by the same group looked at email specifically, and found that batching email into set windows left people measurably less stressed at equal productivity. Which tells you the problem is not volume on its own. It is that the volume arrives continuously.
This is the part most inbox advice gets backwards. Reading faster does not help, because reading was never the expensive part. The expensive part is the decision you make on arrival, several hundred times a day.
What follows from this
If interruption is the cost rather than reading time, then the useful interventions are the ones that reduce the number of decisions, not the ones that speed up the reading.
That means fewer things reaching you unsorted. It means a message that needs nothing from you never becoming a decision at all. And it means the things you agreed to being tracked somewhere other than your memory of a thread, since the alternative is checking, which is another interruption you scheduled for yourself.
Folders do not do this, because filing happens after the decision. Neither does a faster client. What helps is anything that arrives at the decision before you do, and shows its reasoning so you can disagree with it.
That is the job Smart Inbox exists to do in CevroFlow: every message sorted and ranked before you open it, with a one-line reason attached to each call, so the inbox hands you a shortlist rather than a queue.
