What the research actually says about email overload
Most “email statistics” pages quote each other, and a few of the best-known numbers trace to no study at all. This is a reading of the primary studies and reports, with the year the data was collected, how many people it covers, and how it was measured.
Read-only, about a minute. Nothing in your mailbox changes.
How much email there is
The global figure is an estimate, and it says little about any one inbox. Statista puts the number of emails sent and received worldwide at about 376 billion a day in 2025, forecast to reach 424 billion in 2026 (Statista, 2025). The page does not name its underlying source or say whether spam is included, and spam is a large share of the total: Kaspersky’s own filters put it at 47.27% of global email traffic in 2024, up 1.27 percentage points on the year before (Kaspersky, 2024 data). Roughly half of all email, in other words, never reaches a person.
The per-person number is more useful. Microsoft’s 2025 Work Trend Index, built on aggregated Microsoft 365 telemetry up to February 2025 and a survey of 31,000 knowledge workers in 31 markets, reports that the average worker receives 117 emails a day, alongside 153 chat messages per weekday (Microsoft, 2025). That is a figure for people in large organisations on one vendor’s platform, so a small business or a personal mailbox may sit well below it, but it is a measured count rather than a survey recollection.
Email also remains the tool people rank first. In Pew’s 2014 survey of 535 employed internet users, 61% said email was “very important” for doing their job, ahead of the internet itself at 54% (Pew Research Center, 2014). The data is a decade old, and Pew noted the sample was not representative of all workers, but no later survey has shown email losing that position.
How much time it takes
The most-quoted figure is the oldest. The McKinsey Global Institute’s 2012 report on social technologies estimated that the average interaction worker spends about 28% of the workweek managing email, and close to 20% more looking for internal information (McKinsey Global Institute, 2012). Two things are worth knowing about it: the data is from 2012, and it is an estimate assembled from other sources rather than a measurement of anyone’s day.
Self-reported figures run higher. Adobe’s 2019 survey of 1,002 U.S. adults found people saying they spent more than three hours a day on work email and more than two on personal email (Adobe, 2019). Self-reports of time are known to be unreliable in both directions, and the survey counts everything from a glance at a phone to an hour of writing.
Measured figures are lower and more specific. When Microsoft looked at intentional activity across its own apps in early 2023, the average employee spent 57% of their time communicating (meetings, email and chat) and 43% creating documents; the heaviest 25% of email users spent 8.8 hours a week on email (Microsoft, 2023). In a 2016 study that logged the computers of 40 information workers for about twelve workdays each, participants averaged 1 hour 23 minutes a day in their email client and opened it 77 times a day on average, with a median of 58 (Mark et al., 2016). The hours vary by job; the checking frequency is the constant.
Interruptions and attention
The interruption studies are smaller, older and more careful than the volume figures. In 2003, researchers at Loughborough recorded the screens of 15 employees at a UK company for 28 working days. Most people’s email client checked for new mail every five minutes, the default; 70% of new emails were opened within six seconds of arriving; and getting back to the previous task at the previous pace took 64 seconds on average (Jackson, Dawson & Wilson, 2003). The recovery is short. The problem is the frequency: the authors recommended checking no more often than every 45 minutes.
Gloria Mark’s group at UC Irvine shadowed 24 information workers for three days each in 2004, timing every switch to the second. People spent about 11 minutes in a project before moving to another, 57% of those working sessions were interrupted, and when interrupted work was resumed the same day, it took 25 minutes 26 seconds on average to get back to it, usually after two or more other activities (Mark, Gonzalez & Harris, 2005).
That last number is the ancestor of the claim that “it takes 23 minutes to refocus after an interruption”. The 23-minute figure does not appear in the 2005 paper or in the group’s 2008 follow-up; it comes from later interviews. And the published figure measures something different from what the claim implies: not how long it takes to regain concentration, but how long passed before someone returned to an interrupted project, with other work done in between. It was also observed in 24 people at one company in 2004.
The controlled version is more surprising. In a 2008 lab experiment with 48 participants answering emails while being interrupted by phone or chat, interrupted work was finished faster with no drop in quality; the cost showed up instead as higher reported stress, frustration, time pressure and effort after about twenty minutes (Mark, Gudith & Klocke, 2008). People compensate for interruptions by speeding up, and pay for it in another currency.
Mark’s longer-run attention figures are from logging studies rather than surveys. In 2004 her group measured average attention on a single screen at about two and a half minutes; across studies from 2016 to 2020 the average was 47 seconds (Mark, 2024). Microsoft’s telemetry points the same way at scale: employees are interrupted by a meeting, email or chat about every two minutes during core hours (Microsoft, 2025).
Stress and checking frequency
Two studies tested cause rather than correlation. In 2012, 13 information workers at a U.S. Army research centre had email removed for five workdays after a baseline week, while wearing heart-rate monitors. Without email, heart-rate variability showed lower stress, and people switched windows about 18 times an hour instead of 37 (Mark, Voida & Cardello, 2012). The sample is tiny and the setting was one workplace, but it is one of the few experiments that changed email and measured the body.
The second is the study behind most “check email three times a day” advice. Kushlev and Dunn had 124 adults spend one week checking email at most three times a day with alerts off, and another week checking as often as they could, in counterbalanced order. In the limited week people reported lower daily stress and felt less distracted by email, while the number of emails they received and answered did not change (Kushlev & Dunn, 2015). The effects were real but modest, around a third of a standard deviation on daily stress, and the authors chose three a day because it was a limit most people could keep, not because three is the right number.
The 2016 logging study complicates the batching advice. More time on email in a day went with higher measured stress and lower self-rated productivity, but people who batched their email were no less stressed than those who checked continually (Mark et al., 2016). Checking less often helped in the experiment; simply grouping the checks did not show the same benefit in the field.
After-hours email is less about the email than the expectation. A 2016 survey study found that an organisation’s expectation that people monitor email outside work hours, not the time they actually spent on it, predicted emotional exhaustion and worse work-family balance; the authors called the mechanism “anticipatory stress” (Belkin, Becker & Conroy, 2016). The same group later replicated this across three samples, including 138 employee-partner pairs, and found the anxiety carried over to partners’ health and relationship satisfaction (Becker et al., 2019). The expectation is common: in 2014, 36% of full-time U.S. workers told Gallup they frequently checked email outside working hours and 33% said their employer expected them to (Gallup, 2014). In 2025, Microsoft’s telemetry showed 29% of active workers back in their inbox by 10 pm, and meetings after 8 pm up 16% on the year before (Microsoft, 2025).
What the research does not show
- A right number of checks a day. One experiment tested three against unlimited and found a modest benefit (Kushlev & Dunn, 2015); one field study recommended 45-minute gaps on the strength of 15 people (Jackson, Dawson & Wilson, 2003). Nothing has compared two against five against ten.
- A fixed cost per interruption. The measured recovery times range from 64 seconds for a glance at email to over 25 minutes before returning to an abandoned project. These measure different things, and neither is “23 minutes to refocus”.
- Hours per day for the average person. The large figures are either self-reported (Adobe, 2019), estimated from other estimates (McKinsey Global Institute, 2012), or drawn from one vendor’s enterprise customers (Microsoft, 2023; Microsoft, 2025). The measured studies are small. A quarter of desk workers say they spend too much time on email, and 43% of executives do (Slack, 2023), which tells you how people feel, not how long they spend.
- That survey figures from vendors are neutral. Microsoft, Adobe, Slack and Kaspersky all sell something that the problem makes more valuable, and so does Syfft. The telemetry-based figures are still the best large-sample counts available; read them as counts of a particular population, not of everyone.
What it means for one inbox
- Check in sessions, and fewer of them. The one experiment that reduced checking reduced stress without reducing the mail handled (Kushlev & Dunn, 2015). Pick a number you can keep; the evidence does not favour a particular one.
- Turn off new-mail alerts. With alerts on, most people open a new email within six seconds (Jackson, Dawson & Wilson, 2003), and the limited-checking condition that worked had alerts off (Kushlev & Dunn, 2015). A session is only a session if nothing interrupts it.
- Separate what needs you from what does not before you read. Interrupted work was resumed the same day only after two or more other activities (Mark, Gonzalez & Harris, 2005). The fewer emails that pull you into a new activity, the fewer resumptions you pay for; a list that shows which messages wait on you keeps the rest from becoming detours.
- Cut inflow at the source. At 117 emails a day (Microsoft, 2025), a session that handles everything is long. Unsubscribing, turning off app notifications and routing automated mail out of the inbox shortens every session you keep. Our guide to too many emails goes through the sources.
- End each session at a known state. The after-hours studies found the expectation of unanswered mail, not the mail itself, was the stressor (Belkin, Becker & Conroy, 2016; Becker et al., 2019). Knowing that nothing in the inbox is waiting on you is what lets a session end. Our guide to email triage describes one way to get there.
Where Syfft fits
Syfft works on the third point. It reads each new email as it arrives in Gmail or Outlook and adds one label for what the email needs from you: Your turn, Their turn, Deadline mentioned, Newsletters, Promotions, Automated updates and the rest of its labels. It moves nothing, so the inbox and its order stay as they were; every label is logged and can be undone; unclear mail gets no label. You start read-only, with the labels shown on your 50 most recent emails before anything is added to your mailbox. It will not make the mail shorter, but it makes the list answer the question the research keeps circling: which of these is actually waiting on me. If the problem feels more like exhaustion than backlog, our guide to email fatigue starts from there.
Sources
- Statista Research Department, “Number of sent and received e-mails per day worldwide”, 2025 estimate (updated September 2025). statista.com
- Kaspersky (Securelist), “Spam and phishing in 2024”, 2024 data from Kaspersky’s own filters, published February 2025. securelist.com
- Microsoft, Work Trend Index 2025, “Breaking down the infinite workday”: Microsoft 365 telemetry to February 2025 and a survey of 31,000 knowledge workers in 31 markets. microsoft.com
- Microsoft, Work Trend Index 2023, “Will AI fix work?”: Microsoft 365 telemetry for the 28 days to March 2023 and a survey of 31,000 workers in 31 markets. microsoft.com
- Pew Research Center, “Technology’s Impact on Workers”, December 2014; 535 employed adults surveyed September 2014. pewresearch.org
- McKinsey Global Institute, “The social economy: Unlocking value and productivity through social technologies”, July 2012; time-use shares are estimates from secondary sources. mckinsey.com
- Adobe, 2019 Email Usage Study, “If you think email is dead, think again”; survey of 1,002 U.S. adults, July 2019. blog.adobe.com
- Gloria Mark, Shamsi Iqbal, Mary Czerwinski, Paul Johns and Akane Sano, “Email Duration, Batching and Self-interruption: Patterns of Email Use on Productivity and Stress”, CHI 2016; 40 information workers logged for about 12 workdays. microsoft.com (PDF)
- Tom Jackson, Ray Dawson and Darren Wilson, “Reducing the effect of email interruptions on employees”, International Journal of Information Management 23(1), 2003; 15 employees recorded over 28 working days. interruptions.net (PDF)
- Gloria Mark, Victor Gonzalez and Justin Harris, “No Task Left Behind? Examining the Nature of Fragmented Work”, CHI 2005; 24 information workers observed for three days each. ics.uci.edu (PDF)
- Gloria Mark, Daniela Gudith and Ulrich Klocke, “The Cost of Interrupted Work: More Speed and Stress”, CHI 2008; lab experiment with 48 participants. interruptions.net (PDF)
- Gloria Mark, interviewed by the University of California newsroom, “How to sharpen your attention and meet your goals”, 2024; the 2004 and 2016–2020 attention figures in her own words. universityofcalifornia.edu
- Gloria Mark, Stephen Voida and Armand Cardello, “A Pace Not Dictated by Electrons: An Empirical Study of Work Without Email”, CHI 2012; 13 information workers, five workdays without email. UC Irvine’s release gives the window-switching figures. news.uci.edu and doi.org
- Kostadin Kushlev and Elizabeth Dunn, “Checking email less frequently reduces stress”, Computers in Human Behavior 43, 2015; 124 adults, two-week within-subjects experiment. doi.org
- Liuba Belkin, William Becker and Samantha Conroy, “Exhausted, but Unable to Disconnect”, Academy of Management annual meeting, 2016; survey of working adults (Lehigh’s release gives the sample as 297 in one place and 385 in another). news.lehigh.edu
- William Becker, Liuba Belkin, Samantha Conroy and Sarah Tuskey, “Killing Me Softly: Organizational E-mail Monitoring Expectations’ Impact on Employee and Significant Other Well-Being”, Journal of Management, 2019; three studies of 108, 138 pairs and 162 U.S. working adults. doi.org
- Gallup, “For Workers, an Upside to Staying Connected to Work”, 2014; Gallup Daily tracking, 3,865 full-time U.S. workers, March–April 2014. news.gallup.com
- Slack, Workforce Index, “The surprising connection between after-hours work and decreased productivity”, December 2023; survey of 10,333 desk workers in six countries, August–September 2023. slack.com
See what Syfft finds in your inbox.
Connect Gmail or Outlook and see the labels on your 50 most recent emails. Read-only, and it takes about a minute.