PeakLook · Research

How much time do we spend on email? 28% of the week, and how to get a day back

How much time do we spend on email? McKinsey measured 28% of the work week; Microsoft puts 57% of time on communication. The two changes that give a day back.

By Hen Eytan, updated 6 min read4 sources
11.2 hHours a week on email, typical4.5 hWith drafting and follow-ups automatedIllustrative: the McKinsey share of a 40-hour week versusreading and deciding only

Key takeaways

  • Knowledge workers spend about 28% of the week on email; Microsoft's 2023 usage data puts 57% of time on communication overall.
  • Longer daily email time went with lower perceived productivity and higher measured stress in a 40-person field study; checking less often lowered stress in a randomised trial.
  • The hours go to drafting, triage and follow-up, not reading.
  • Batch the inbox and automate the mechanical parts; keep the deciding.

Email is the job that surrounds the job. The McKinsey Global Institute put a number on it that has not moved much in a decade, Microsoft's own telemetry added a second one in 2023, and a Vancouver lab put a number on what it costs you to keep checking. Here is what the research says about how much time we spend on email, where the hours go, and the two changes that give most of a day back.

How much time do we spend on email each week?

MGI's 2012 Social Economy report estimated that the average interaction worker spends 28% of the workweek reading and answering email, and another 19% searching for information. That is more than eleven hours in a forty-hour week. Later surveys have found similar or larger figures, and the arrival of chat tools added to the total rather than replacing it.

Microsoft's 2023 Work Trend Index measured it differently, from Microsoft 365 usage data rather than a survey: the average employee spent 57% of their time communicating (meetings, email and chat) and 43% creating (documents, spreadsheets, presentations). Email alone took 15% of the week on that chart, and the heaviest quarter of email users spent 8.8 hours a week inside it. The survey half of the same report polled 31,000 workers in 31 markets.

SourceWhat it measuredFigure
McKinsey Global Institute, 2012Share of the workweek spent reading and answering email28%
McKinsey Global Institute, 2012Share spent searching for information19%
Microsoft Work Trend Index, 2023Share of time spent communicating versus creating57% versus 43%
Microsoft Work Trend Index, 2023Hours a week on email, heaviest 25% of users8.8 h

The two measurements disagree on the exact share and agree on the shape: communication, not production, is where most of a knowledge worker's week goes, and email is the largest fixed slice of it.

What does checking email all day cost?

Kushlev and Dunn randomised 124 adults to check email only three times a day for a week, and freely the other week. Stress was significantly lower in the limited week. The mechanism is the one Gloria Mark documented in office workers: an interruption costs about 23 minutes to fully recover from, and an open inbox is a machine for producing interruptions.

A later field study by Mark and colleagues at Microsoft Research logged 40 information workers for 12 workdays with computer logging, biosensors and daily surveys. The longer someone spent on email each day, the lower their perceived productivity and the higher their measured stress. Duration itself is the cost; the share of the week is not an abstraction. We go through the interruption evidence in detail in how often you should check email.

PeakLook does this from your own data. Drafts the reply. Keeps the follow-up going. Join the free beta →

Where do the hours actually go?

Not reading. Reading is fast. The hours go to drafting replies, deciding what needs a reply at all, and remembering to follow up on the ones that went unanswered. Those three tasks are the ones a machine can do well: drafting is language, triage is classification, and follow-up is a schedule.

  • Drafting. The blank page is the slow part. A reply that takes eight minutes to write from nothing takes two to edit from a decent draft.
  • Triage. Newsletters, receipts and notifications are the majority of arrivals and need no decision at all, yet each one gets looked at.
  • Chasing. The thread you sent on Monday with no reply has no trigger to remind you on Thursday. Most people simply do not follow up, and the ask dies.

We map which of these an assistant should own, and which it should not, in what an AI email assistant can and cannot do.

Is inbox zero the answer?

Only in its original sense. Merlin Mann's point was never an empty folder; it was how much of your attention the inbox holds. A decided inbox is the goal: every thread either has a reply drafted, a follow-up scheduled, or a reason to be ignored. An inbox that is empty because everything was archived unread is not decided, and an inbox with forty threads that are all waiting on someone else is fine. The inbox zero method is a set of decisions, not a count.

How to get a day back: two changes

Two changes carry most of the benefit. Batch: open the inbox two or three times a day and treat it as a task, not a feed. Automate the mechanical parts: let a drafter write the first version of every reply, let a sequence carry the follow-ups, and spend your attention on reading and deciding.

Inbox taskWho should do itWhat changes
Reading and decidingYouNothing; this is the job
Writing the first draft of a replyAn assistant, in your voiceYou edit instead of starting from a blank page
Sorting newsletters and notifications outFilters, or an assistantThey never reach the batch
Chasing threads with no replyA scheduled sequenceFollow-ups go out on day 3, 7 and 14 without memory
Checking for new mailA schedule, not a notificationTwo or three sessions a day

The same arithmetic applies to any interrupt-driven job; the developer version is in interruptions cost developers 23 minutes. The point is not to answer faster. It is to make the checking a scheduled task that is mostly done by the time you open the tab.

How PeakLook gives the day back

PeakLook is built for the second change. It scans the Focused inbox in Microsoft 365, skips the newsletters, and saves a Reply-All draft threaded to every conversation that needs one. Follow-up sequences run server-side on a daily schedule with no window open, and an enrolment halts the instant a reply lands. Anything the machine wrote or will send is shown in gold, so you always know who wrote what. What is left when you open the inbox is reading and deciding, which is the part only you can do.

Want this answered from your own data?

That is what PeakLook does. Drafts the reply. Keeps the follow-up going. In private beta — the waitlist is free.

Join the beta About PeakLook

Questions

How many hours a week does the average person spend on email?
The McKinsey estimate is 28% of a working week, about eleven hours of forty. Microsoft's 2023 Exchange Online data found the heaviest quarter of users spent 8.8 hours a week on email alone, and communication of all kinds took 57% of the average employee's time.
Is inbox zero the goal?
No. A decided inbox is. Every thread either has a reply drafted, a follow-up scheduled, or a reason to be ignored. An empty inbox you reached by archiving everything unread has decided nothing.
Does AI drafting produce generic email?
It produces a first draft in your voice with the thread's context; you edit and send. The time saved is the blank page, not the judgement, and the judgement stays yours.
How many times a day should I check email?
Kushlev and Dunn used three. Two to four works for most roles; what matters is that it is a scheduled task with a small batch, not a feed that is always open.
Does spending less time on email hurt productivity?
The evidence points the other way. In Mark's 2016 field study of 40 information workers, longer daily email duration went with lower perceived productivity and higher stress, and batching went with higher rated productivity. Less time in the inbox, in short sessions, is the pattern the data favours.

Sources

  1. McKinsey Global Institute (2012). The social economy: unlocking value and productivity through social technologies. https://www.mckinsey.com/industries/technology-media-and-telecommunications/our-insights/the-social-economy
  2. Microsoft (2023). Work Trend Index Annual Report: Will AI fix work? https://www.microsoft.com/en-us/worklab/work-trend-index/will-ai-fix-work
  3. Kushlev K., Dunn E.W. (2015). Checking email less frequently reduces stress. Comput Human Behav. https://www.sciencedirect.com/science/article/abs/pii/S0747563214005810
  4. Mark G., Iqbal S., Czerwinski M., Johns P., Sano A. (2016). Email duration, batching and self-interruption: patterns of email use on productivity and stress. CHI. https://www.microsoft.com/en-us/research/publication/email-duration-batching-and-self-interruption-patterns-of-email-use-on-productivity-and-stress/
Hen Eytan · Founder, UPeak

Builds the UPeak apps and writes up the research behind them.