Skip to content
WriteVoice WriteVoice Start Creating

143 Email Statistics October 2026

Updated 143 primary-source citations Refreshed monthly

On this page
  1. Key Takeaways
  2. How many people use email?
  3. How many emails are sent per day?
  4. How many emails does the average worker receive per day?
  5. How much time do people spend on email at work?
  6. How much email happens outside working hours?
  7. How much effort does writing an email take?
  8. Methodology and Sources

Email statistics show a channel that is still growing. Microsoft alone screens 5.2 billion emails a day on average for malware and phishing (Microsoft Digital Defense Report 2026, Oct 2026).

Most written work still runs through email, and reading and answering it now spills into early mornings, evenings and phones. The data below runs from who uses email and how much is sent to how long it takes and how much of the writing software already does.

Every number links to the organization that produced it and carries the date of its data. Telemetry, computer logs, lab studies and surveys measure different things, so each figure is labelled with its method. The page is refreshed monthly.

Key Takeaways

How many people use email?

More adults in Canada and the EU use email than any social network, and the people left out are mostly older, on low incomes or offline.

Email statistics: share of people aged 16 to 74 who sent or received email in 2025 in each EU country, next to calls, social networks and instant messaging, from 98% in the Netherlands down to 46% in Romania, with seven non-EU countries listed below, from 96% in Norway to 41% in North Macedonia
Figure 1: Email is one of the two most common ways Europeans communicate online and runs well ahead of social networks, but the spread between countries is wide, from near-universal use in the Netherlands and Denmark to a minority of adults in Romania. The table lists the EU countries first and non-EU countries below them. Source: Eurostat, Digital economy and society statistics, households and individuals, Dec 2025. Table published by the source, captured 2 Oct 2026.

How many email users are there worldwide?

Billions, by the only estimate available. No government counts email users worldwide, so the number comes from one research firm, and the internet population sets the ceiling.

  • The Radicati Group counted 4.48 billion business and consumer email users worldwide in 2024 and forecast 4.73 billion for 2026 and 4.97 billion by 2028, growth of about 3% a year; the 2026 figure is a Dec 2024 forecast, not a count (Radicati Group, Email Statistics Report 2024-2028 (archived copy), Dec 2024).
  • An estimated 6 billion people, about three-quarters of the world's population, used the internet in 2025, up from 5.8 billion in 2024 (ITU, Facts and Figures 2025, Nov 2025).
  • 2.2 billion people were still offline in 2025, and 96% of them live in low- and middle-income countries (ITU, Facts and Figures 2025, Nov 2025).
  • 94% of people in high-income countries use the internet, against 23% in low-income countries, so nearly all the room left for new email users is in poorer countries (ITU, Facts and Figures 2025, Nov 2025).
  • Google's own engineers wrote in 2019 that Gmail was used by more than 1.5 billion people (Chen et al., KDD 2019, Aug 2019).

Who uses email, by age and country?

Almost every working-age adult who is online. Government email usage statistics from the US, Canada and the EU agree on the pattern: use is near-universal among employed people aged 25 to 44 and drops for people over 65, for low-income households and in a few Eastern European countries. NTIA counts US internet users aged 15 and older, Statistics Canada counts all Canadians aged 15 and older, and Eurostat counts all people aged 16 to 74, so the three rates are not directly comparable.

  • Email use among US internet users stood at 83.7% in Jul 2013, about 9 points below its Nov 2023 level (NTIA Data Explorer, Nov 2023).
  • Americans aged 25 to 44 are the heaviest email users at 95.0% of internet users, ahead of 93.1% of those aged 15 to 24 and 86.5% of those 65 and older (NTIA Data Explorer, Nov 2023).
  • 95.7% of employed US internet users use email, against 86.2% of those not in the labor force (NTIA Data Explorer, Nov 2023).
  • Income splits email use more than age does: 97.3% of internet users in households earning $100,000 or more use email, against 83.4% in households under $25,000 (NTIA Data Explorer, Nov 2023).
  • The Netherlands (98%), Denmark (97%), Ireland and Finland (both 95%) had the highest email use in the EU in 2025 (Eurostat, Digital Economy and Society Statistics, Dec 2025).
  • Fewer than half of adults in Romania (46%) sent or received email in 2025, and Bulgaria was at 52%, even though about three-quarters of adults in both countries made online calls (Eurostat, Digital Economy and Society Statistics, Dec 2025).
  • 89.6% of Canadians aged 15 and older sent and received email in 2022, up from 85.9% in 2020 and 85.5% in 2018 (Statistics Canada, Canadian Internet Use Survey, Jul 2023).
  • Email was the most common of the online activities in Statistics Canada's 2022 table, ahead of instant messaging (78.0%) and social networking (73.7%) (Statistics Canada, Canadian Internet Use Survey, Jul 2023).
  • 96.2% of Canadians aged 25 to 44 used email in 2022, against 74.7% of those 65 and older (Statistics Canada, Canadian Internet Use Survey, Jul 2023).

Which apps and devices do people read email on?

Mostly phones, and mostly Apple Mail and Gmail. Litmus counts opens of the marketing email its customers track, not people, and Apple's Mail Privacy Protection inflates Apple's share, so read its numbers as a guide to which apps open email rather than a census of users.

  • Apple Mail on iPhone, iPad and Mac accounted for 62.26% of more than 1 billion email opens Litmus tracked in Jul 2026 (Litmus Email Client Market Share, Jul 2026).
  • Gmail accounted for 27.03% of tracked opens in Jul 2026, desktop Outlook for 5.83% and Yahoo Mail for 2.59% (Litmus Email Client Market Share, Jul 2026).
  • Almost 9 in 10 EU internet users went online with a mobile phone in 2025, against 56% on a laptop and 31% each on a desktop computer or tablet (Eurostat, Digital Economy and Society Statistics, Dec 2025).
  • 91% of US adults own a smartphone, up from 35% when Pew first asked in 2011 (Pew Research Center, Mobile Fact Sheet, Nov 2025).
  • Smartphone ownership falls from 97% of US adults aged 18 to 29 to 78% of those 65 and older, the same age group that uses email least in the NTIA data above (Pew Research Center, Mobile Fact Sheet, Nov 2025).
  • 16% of US adults are smartphone-only internet users: they own a smartphone but have no home broadband, so their inbox lives on their phone (Pew Research Center, Mobile Fact Sheet, Nov 2025).
  • In a 2015 study of 2 million Yahoo Mail users, replies sent from phones came back in a median of 28 minutes, against 57 minutes from tablets and 62 minutes from desktops (Kooti et al., WWW 2015, May 2015).
  • Phone replies in the same study had a median length of 20 words, against 27 words from tablets and 60 words from desktops (Kooti et al., WWW 2015, May 2015).
  • Users aged 35 to 50 were the most likely to have sent email from a phone or tablet in that study (53%) and users over 50 the least likely (43%); women were more likely than men (50% vs 45%) (Kooti et al., WWW 2015, May 2015).

How many emails are sent per day?

Daily email traffic is still rising while paper mail shrinks, and close to half of what is sent is mail nobody asked for. The email volume growth figure everyone quotes is one research firm's forecast. The spam share comes from the security labs that filter the mail, and it is the sturdier number.

Share of spam in global email traffic by month in 2025: between 42.35% and 45.01% from January to October, then 52.87% in November and 51.80% in December
Figure 2: Spam held steady for ten months of 2025 and then jumped in November and December, so the busiest shopping weeks of the year were also the busiest weeks for junk mail. Source: Kaspersky, Spam and phishing in 2025, Feb 2026. Chart published by the source, captured 2 Oct 2026.

How many emails are sent and received each day worldwide?

Hundreds of billions. No public body measures global email traffic; the one estimate comes from The Radicati Group, and its figures after 2024 are projections. Paper mail, which the US Postal Service does count, is moving the other way.

How many spam emails are sent per day?

Billions every day. Gmail alone blocks billions of unwanted messages daily, and the labs that filter mail class close to half of all global traffic as spam. In 2025, for the first time in several years, the fourth quarter was the busiest.

How much does email fraud cost?

Billions of dollars a year in reported losses, most of it through business email compromise, where a fake executive or supplier asks for a payment. Phishing is the crime people report most to the FBI, and the first message in an impersonation scam now rarely asks for anything.

  • Business email compromise losses reported to the FBI reached $3.05 billion in 2025, up 10% from 2024 (FBI IC3, 2025 Internet Crime Report, Apr 2026).
  • The FBI received 24,768 business email compromise complaints in 2025, up from 21,442 in 2024 (FBI IC3, 2025 Internet Crime Report, Apr 2026).
  • Phishing and spoofing was the most reported crime type to the FBI in 2025 with 191,561 complaints, more than twice the next category, extortion, at 89,129 (FBI IC3, 2025 Internet Crime Report, Apr 2026).
  • Only 3% of first-contact impersonation emails Microsoft tracked asked for money or documents by Jun 2026, down from 17% in Oct 2025: the opening message is a bland check-in, and the fraud request comes later in the thread (Microsoft Digital Defense Report 2026, full report, Oct 2026).

How many emails does the average worker receive per day?

Workers get more email than they can answer, and they cope by skimming and by replying to a smaller share of what arrives.

Two line charts from a 2015 Yahoo Mail study plotting the share of received emails that get a reply against emails received per day, from 0 to 100. For all users the share falls from about 0.25 at low volume to under 0.05 at 100 emails a day; low-activity users fall faster; the right panel, counting only mail from known contacts, shows the same decline
Figure 3: As a Yahoo Mail user's daily inbox grows, the share of messages they answer keeps falling, and the decline holds even when only mail from known contacts is counted. Source: Kooti et al., WWW 2015, Evolution of Conversations in the Age of Email Overload, May 2015. Chart published by the source, captured 2 Oct 2026.

How many emails do office workers get each day?

The average Microsoft 365 user receives 117 emails a day, and most of them are skimmed in under 60 seconds, based on Microsoft telemetry that excludes EU and education tenants (Microsoft Work Trend Index, Breaking Down the Infinite Workday, Jun 2025). The average emails received per day has been measured from real mailboxes since the 2000s, and more of the mail now goes to large lists, with chat arriving on top. The meetings that fill the rest of the calendar are covered in our meeting statistics.

What happens to replies as email volume rises?

People answer a smaller share of it. In a study of 16 billion Yahoo Mail messages, the busier the day, the lower the share of email that got a reply, and inboxes and unread piles grew in every study from 1996 to 2012.

  • Across more than 2 million Yahoo Mail users and 16 billion emails, people replied to about 25% of the emails they received on light days and to less than 5% on days with about 100 incoming emails (Kooti et al., WWW 2015, May 2015).
  • Counting only emails from contacts the user had written to, the share answered still fell as daily volume rose (Kooti et al., WWW 2015, May 2015).
  • The most active emailers coped better: under heavy load they sent replies about twice as long as low-activity users, whose replies shrank as their inboxes filled (Kooti et al., WWW 2015, May 2015).
  • Teenagers replied to a near-constant share of their email as volume rose, while the answered share fell further for each older age group (Kooti et al., WWW 2015, May 2015).
  • Most email conversations are short: the median thread held 2 emails (mean 3.76), half of all threads ended within 3.5 hours, and 22% ran longer than a day (Kooti et al., WWW 2015, May 2015).
  • Email archives grew tenfold in a decade, from an average of 2,482 messages in 1996 to 28,660 in 2006, while the average inbox held 1,150 messages (median 512) (Fisher et al., CSCW 2006, Nov 2006).
  • The same 2006 Microsoft mailboxes held a mean of 153 unread messages, although the median worker had only 7, so a minority sat on very large unread piles (Fisher et al., CSCW 2006, Nov 2006).
  • In a 2012 study of 19 US Gmail users, work inboxes held an average of 3,003 emails (median 1,483), nearly twice the 1996 average of 1,624 and 2.6 times the 2006 average of 1,150, and personal inboxes averaged 15,030 (Grevet et al., CHI 2014, Apr 2014).
  • Unread work email in that Gmail sample averaged 696 messages, more than four times the 2006 mean, while personal accounts averaged 4,846 unread (median 421) (Grevet et al., CHI 2014, Apr 2014).

Which jobs carry the heaviest inbox?

Primary care physicians are the best-measured case. Their electronic health record inbox, an email-like queue of patient messages, test results and refill requests, takes a large part of each workday, and patient messages have stayed above their pre-2020 level. These figures describe workload only; the wellbeing data is in our physician burnout statistics.

  • Primary care physicians spent 5.9 hours of an 11.4-hour workday in the electronic health record, including 85 minutes (23.7% of EHR time) on inbox management (Arndt et al., Annals of Family Medicine 2017, Sep 2017).
  • Among 1,275 primary care physicians in a large US medical group, those with the most inbox time viewed 200 messages per workday, against 109 for those with the least, and spent 30 minutes on the inbox outside work hours against 9.7 (Akbar et al., JAMIA, Oct 2020).
  • Among 141 academic primary care physicians, inbox time per 8 hours of scheduled clinic rose 24.4% (14 minutes) between 2019-20 and 2022-23 (Arndt et al., Annals of Family Medicine 2024, Jan 2024).
  • Over the same period, patient medical advice messages to those physicians rose 55.5% (5.4 more per 8 clinic hours), while patient phone calls fell 10.5% (Arndt et al., Annals of Family Medicine 2024, Jan 2024).
  • In national Epic data covering 280,712 US ambulatory physicians, primary care physicians received a mean of 16 patient medical advice messages and 24 patient calls a week, and message volume rose in Mar 2020 and did not return to pre-pandemic levels through early 2022 (Holmgren et al., JAMA Internal Medicine, Feb 2025).
  • Primary care physicians' active EHR time rose 6.5%, from 10.6 to 11.3 hours a week, after the post-2020 jump in patient messages (Holmgren et al., JAMA Internal Medicine, Feb 2025).
  • EHR logs from 471 primary care physicians in 2011-14 showed 3.17 hours a day on desktop medicine, which includes patient portal and staff messages, slightly more than the 3.08 hours spent on office visits (Tai-Seale et al., Health Affairs, Apr 2017).

How much time do people spend on email at work?

Email takes a large slice of the working day under every method used to measure it. Logged data comes in lower than surveys, and both put it ahead of much of the work people are hired to do. Several of the figures below are Microsoft telemetry, and Microsoft also sells AI tools for email.

Share of time spent in Microsoft 365 apps, commercial users, Mar 2023: Teams meetings 23%, Teams chat 19% and email 15% make up 57% communicating; Excel 18%, Word 10%, PowerPoint 8% and OneNote 7% make up 43% creating
Figure 4: Email sits on the communicating side of the scale with meetings and chat, and together they outweigh the documents, spreadsheets and decks where the work gets made. Source: Microsoft Work Trend Index, Will AI Fix Work?, May 2023. Screenshot of the source page, captured 2 Oct 2026.

How many hours a week go to email?

Several hours for most office workers, and more than a full workday for the heaviest users. The figures come from Microsoft telemetry, computer logging and one consulting estimate, so read each with its method.

  • The average employee spends 57% of their Microsoft 365 time communicating in meetings, email and chat, and 43% creating in documents, spreadsheets and presentations, based on telemetry from commercial customers (Microsoft Work Trend Index 2023, May 2023).
  • Email alone takes 15% of that Microsoft 365 time, behind Teams meetings at 23% and Teams chat at 19% (Microsoft Work Trend Index 2023, May 2023).
  • 68% of workers say they do not have enough uninterrupted focus time during the workday, in a survey of 31,000 full-time workers in 31 markets (Microsoft Work Trend Index 2023, May 2023).
  • The most-pinged fifth of Microsoft 365 users are interrupted by a meeting, email or chat every 2 minutes during core work hours, 275 times across a 24-hour day (Microsoft Work Trend Index, Breaking Down the Infinite Workday, Jun 2025).
  • McKinsey Global Institute estimated in 2012 that the average interaction worker, a high-skill knowledge worker such as a manager or professional, spends 28% of the workweek managing email (McKinsey Global Institute, The Social Economy, Jul 2012).
  • In a 2016 study that logged the computers of 40 information workers for about 12 workdays, participants averaged 1 hour 23 minutes a day in email out of 4 hours 34 minutes of logged computer time (Mark et al., CHI 2016, May 2016).
  • 35% of US online workers said in 2014 that the internet, email and cell phones had increased the number of hours they work (Pew Research Center, Technology's Impact on Workers, Dec 2014).

Does time on email hurt productivity and stress?

In logged field studies, more time on email went with lower self-rated productivity and higher measured stress. These are findings from small samples of office workers, and they describe associations and short experiments, not a diagnosis.

  • The more time a worker spent on email in a day, the lower they rated their productivity that day, and the longer they spent on email in a given hour, the higher their stress measured by heart rate variability in that hour (Mark et al., CHI 2016, May 2016).
  • Batching email into a few blocks a day was linked to higher self-rated productivity at high email loads, but the study found no evidence that batching lowered stress (Mark et al., CHI 2016, May 2016).
  • 41.0% of those workers mainly opened email on their own initiative, 30.8% mainly when a notification arrived and 28.2% about equally both ways; 34 of the 40 had email notifications switched on (Mark et al., CHI 2016, May 2016).
  • In a 2012 UC Irvine and US Army study, office workers with email switched computer windows an average of 37 times an hour; during five days without email they switched about 18 times an hour (UC Irvine News, May 2012).
  • In the same study, 13 information workers had email cut off for five workdays, and heart rate variability pointed to lower stress without email than on their normal working days, a result based on 7 participants with usable heart data (Mark, Voida and Cardello, CHI 2012, May 2012).
  • When 124 adults were randomly assigned to check email only three times a day for one week, they reported significantly lower daily stress than in a week of unlimited checking (Kushlev and Dunn, Computers in Human Behavior, Feb 2015).
  • Their email load stayed the same in both weeks, about 16 emails received and 5 to 6 answered a day, so the difference came from checking less, not from getting less email (Kushlev and Dunn, Computers in Human Behavior, Feb 2015).
  • A 2007 Microsoft Research field study that logged 27 employees found email alerts interrupted their main task 4.28 times an hour, and 40.8% of alerts got a response within 15 seconds (Iqbal and Horvitz, CHI 2007, Apr 2007).
  • After switching to email from an alert, those employees took an average of 9 minutes 33 seconds before returning to any of the tasks they had left (Iqbal and Horvitz, CHI 2007, Apr 2007).

Why do surveys and logs disagree on email time?

Surveys ask people to estimate and logs count what happened. Time spent on email statistics come from three methods: telemetry, computer logs and self-reported surveys. People tend to overestimate the hours and underestimate how often they check. Trust logged data for time and frequency, and surveys for how email feels.

  • US adults told Adobe in 2019 that they spend about five hours a day on email, more than three hours on work email and more than two on personal email, in a self-reported survey of 1,002 people (Adobe Email Usage Study, Sep 2019).
  • Logged email time varies far more than survey averages suggest: across the 40 logged workers in the 2016 study, daily email time ranged from zero to 7 hours 54 minutes, with a median of 1 hour 6 minutes (Mark et al., CHI 2016, May 2016).
  • Asked before a 2015 experiment, participants estimated they checked email 15.48 times on a normal workday (Kushlev and Dunn, Computers in Human Behavior, Feb 2015).
  • In the 2014 Pew survey, 46% of US online workers said the internet, email and cell phones made them more productive and 7% said less productive (Pew Research Center, Technology's Impact on Workers, Dec 2014).

How much email happens outside working hours?

Work messages now reach most workers before and after their working hours. The pressure comes less from the volume than from the sense that every message needs an answer now, and several governments have written a right to disconnect into law.

Share of workers contacted for work-related reasons outside their contractual hours in a typical week, by who contacts them: line managers about 18% every day and 49% some days, colleagues 31% every day and 43% some days, clients 23% every day and 29% some days
Figure 5: Colleagues, not bosses, are the most frequent source of out-of-hours contact, and clients reach the fewest workers after hours. Source: Eurofound, Right to Disconnect: Implementation and Impact at Company Level, Figure 5, Nov 2023. Chart published by the source, captured 2 Oct 2026.

How early and how late do people check work email?

Before the workday starts and well after it ends. Managing email overload is now a question of hours as much as volume: telemetry shows inboxes busy before the workday, late in the evening and on weekends, and surveys in the US and Europe find most workers handle work messages outside their contracted time.

How fast do people expect a reply?

Faster than senders usually need. Measured reply times are short, and experiments show that people who get a non-urgent email after hours overestimate how quickly the sender wants an answer.

  • The single most common reply time was two minutes, and more than 90% of replies came within a day, across more than 2 million Yahoo Mail users and 16 billion emails (Kooti et al., WWW 2015, May 2015).
  • Teenagers replied to email in a median 13 minutes, against 16 minutes for users aged 20 to 35, 24 minutes for users aged 36 to 50 and 47 minutes for users aged 51 and over (Kooti et al., WWW 2015, May 2015).
  • Men replied in a median 24 minutes and women in 28 minutes (Kooti et al., WWW 2015, May 2015).
  • For a reply of the same 50 words, teenagers took a median 39 minutes and users aged 51 and over took 79 minutes (Kooti et al., WWW 2015, May 2015).
  • Email received on weekends got slower replies than weekday email, about 55 minutes against about 45, and email received around 4 am waited roughly four hours for a reply, as read from the paper's Figure 10 (Kooti et al., WWW 2015, May 2015).
  • In eight pre-registered experiments with 4,004 participants, people who received non-urgent work email in the evening or on the weekend overestimated how quickly the sender expected a reply, and the gap was linked to more stress and lower well-being (Giurge and Bohns, OBHDP 2021, Nov 2021).
  • 75% of workers who answer out-of-hours messages say they do it because it is expected, and 61% because they fear a negative impact if they do not reply (Eurofound, Right to Disconnect: Implementation and Impact at Company Level, Nov 2023).
  • In two studies, of 108 US working adults and of 138 employees with their partners, an employer's expectation that staff monitor email outside work hours was linked to higher employee anxiety and to worse health and relationship satisfaction for both the employee and their partner (Becker et al., Journal of Management 2021, author manuscript, Apr 2021). The published version is at doi:10.1177/0149206319890655.

Which countries limit after-hours work email?

Australia, Ontario, Portugal, Spain and France have some form of right to disconnect in law, and Ireland has a statutory code of practice. Australia's rule is the broadest, and the EU has no law yet although the European Parliament asked for one.

  • Australian employees can refuse to monitor, read or respond to contact outside working hours unless the refusal is unreasonable: the right applied from 26 Aug 2024 at employers with 15 or more employees and from 26 Aug 2025 at small businesses with fewer than 15 (Fair Work Ombudsman, Right to Disconnect, Jun 2026).
  • Ontario employers with 25 or more employees must have a written policy on disconnecting from work; the rule was added to the Employment Standards Act on 2 Dec 2021, and the first policies were due by 2 Jun 2022 (Government of Ontario, Written Policy on Disconnecting from Work, Jul 2024).
  • Portugal's Law 83/2021 of 6 Dec 2021 obliges employers to refrain from contacting workers during their rest period except in cases of force majeure, makes a breach a serious offence, and took effect on 1 Jan 2022 (Diario da Republica, Lei 83/2021, Dec 2021).
  • Spain made digital disconnection a workplace right in Article 88 of Organic Law 3/2018, in force since 7 Dec 2018 (BOE, Ley Organica 3/2018, Dec 2018).
  • In France the right to disconnect is one of the topics of mandatory annual bargaining under Labour Code Article L2242-17, and where no agreement is reached the employer must draw up a charter on it after consulting the works council; the current version has been in force since 31 Mar 2022 (Legifrance, Code du travail, Mar 2022).
  • Ireland's Workplace Relations Commission issued a Code of Practice on the Right to Disconnect in Mar 2021 after a public consultation that drew 37 submissions; breaching the code is not an offence, but courts and the WRC must take it into account (Workplace Relations Commission, Code of Practice on the Right to Disconnect, Mar 2021).
  • The European Parliament called for an EU law on the right to disconnect on 21 Jan 2021, by 472 votes to 126 with 83 abstentions (European Parliament, Jan 2021).
  • Around 45% of employees Eurofound surveyed said their company had a right-to-disconnect policy, and workers in those companies reported good work-life balance more often, 92% against 80% (Eurofound, Right to Disconnect: Implementation and Impact at Company Level, Nov 2023).

How much effort does writing an email take?

Most replies are short, and software already drafts a measurable share of them. The strongest evidence that this saves time comes from a randomized trial, not a vendor survey.

Weekly change in hours spent in Outlook email for workers given Microsoft 365 Copilot versus a control group, by month before and after access, across 66 firms: flat near zero before access, then an intention-to-treat (ITT) estimate about 1 to 1.8 hours a week lower and a local average treatment effect (LATE) estimate about 1.6 to 2.7 hours lower in every month after access
Figure 6: The drop in email time showed up in the first month workers had the tool and did not fade over the months the trial measured. Source: Dillon et al., NBER w33795, Shifting Work Patterns with Generative AI, Figure 2(a), May 2025. Chart published by the source, captured 2 Oct 2026.

How long are email replies, and how fast can people write them?

Short, and longer for older writers. The bullets give reply lengths from one study of 2 million mail users and entry speeds from a lab test and two online typing tests, each measured its own way. The full speed data is in our typing speed statistics, and our guide to dictating emails covers the speech side.

  • Reply length grew with age in a 2015 study of 2 million Yahoo Mail users: teenagers' median reply was 17 words, young adults' 21, adults' 31 and users aged 51 and older 40 words (Kooti et al., WWW 2015, May 2015).
  • Men and women wrote replies of almost the same length, a median of 28 words for men and 30 for women (Kooti et al., WWW 2015, May 2015).
  • Long replies are a minority but not rare: 30% were longer than 100 words (Kooti et al., WWW 2015, May 2015).
  • In a lab test on an iPhone, speech recognition entered English text 2.93 times faster than the built-in keyboard, 153 words per minute against 52 (Ruan et al., IMWUT 2017, Dec 2017).
  • 168,000 volunteers who took an online typing test on physical keyboards averaged 51.56 words per minute (Dhakal et al., CHI 2018, Apr 2018).
  • On their own phones, 37,370 volunteers in the same kind of online test averaged 36.17 words per minute (Palin et al., MobileHCI 2019, Oct 2019).

How much email writing do AI tools already do?

A real share already. Google's suggestion tools were assisting with a measurable share of mobile replies by 2016, and Microsoft and Google now report AI assists in the billions. Use among individual workers is wide but shallow: most US workers still rarely or never use AI chatbots on the job.

  • Google's Smart Reply, which offers three one-tap responses, was assisting with 10% of all mobile replies in Inbox by Gmail by mid-2016 (Kannan et al., KDD 2016, Jun 2016).
  • By May 2017 that share had grown to 12% of replies sent from Inbox on mobile, and Google brought Smart Reply to the Gmail apps for Android and iOS (Google, Save Time with Smart Reply in Gmail, May 2017).
  • Gemini in Google Workspace, which includes drafting in Gmail, was giving business users more than 2 billion AI assists a month across Workspace in Apr 2025 (Google Workspace Blog, Apr 2025).
  • Microsoft 365 Copilot, the AI assistant built into Outlook, Word and Teams, passed 30 million paid seats in Microsoft's fiscal year ended Jun 2026 (Microsoft FY26 Q4 Earnings Release, Jul 2026).
  • 75% of knowledge workers surveyed across 31 countries in early 2024 said they used generative AI at work, and 46% of them had started in the previous six months, in a Microsoft and LinkedIn survey (Microsoft Work Trend Index 2024, May 2024).
  • Only about one in ten US workers used AI chatbots such as ChatGPT, Gemini or Copilot at work every day or a few times a week in late 2024, and 55% rarely or never did (Pew Research Center, U.S. Workers and AI, Feb 2025).
  • Among US workers who had used AI chatbots for their job, 52% had used them to edit written content and 47% to draft it, behind only research at 57% (Pew Research Center, U.S. Workers and AI, Feb 2025).
  • In a UK government trial of Microsoft 365 Copilot with 20,000 civil servants, 33% of survey respondents used it daily in Outlook, nearly level with Teams at 34% and ahead of Word at 25% (UK Government Digital Service, M365 Copilot Experiment, Jun 2025).

Does AI save measurable time on email?

Yes, and the strongest evidence is a randomized trial rather than a vendor survey: workers given an AI assistant inside their email spent less time in it with no drop in the threads they answered. Three of the trial's four authors worked at Microsoft Research, and Microsoft makes the tool that was tested. Self-reported savings in surveys run higher and should be read as perceptions.

  • Counting everyone offered the tool, whether they used it or not, email time fell by 1.4 hours a week, 12% of the 11.7 hours the average worker spent reading and answering email before the trial (Dillon et al., NBER w33795, May 2025).
  • Workers with the assistant replied to the same number of email threads as the control group while reading 4% fewer emails, so the saved time did not come from ignoring the inbox (Dillon et al., NBER w33795, May 2025).
  • Workers with access also cut their weekly Outlook sessions by 11% and gained almost 1.5 hours a week in email-free blocks of 30 minutes or more during the regular workday (Dillon et al., NBER w33795, May 2025).
  • In Microsoft's own survey of 297 early Copilot customers in late 2023, 64% said it helped them spend less time processing email and 85% said it got them to a good first draft faster (Microsoft WorkLab, Copilot's Earliest Users, Nov 2023).
  • A blind panel of 62 people rated emails written with Copilot 18% clearer and 19% more concise than emails written without it, in a Microsoft test (Microsoft WorkLab, Copilot's Earliest Users, Nov 2023).
  • UK civil servants in the 20,000-person Copilot trial said they saved an average of 26 minutes a day, which the government puts at 13 working days a year if sustained (UK Government Digital Service, M365 Copilot Experiment, Jun 2025).
  • Only 17% of trial participants noticed no clear time savings, while more than a third said they saved over half an hour a day (UK Government Digital Service, M365 Copilot Experiment, Jun 2025).
  • 40% of US workers who had used AI chatbots for work said the tools were extremely or very helpful for doing things more quickly, against 29% who said the same about improving the quality of their work (Pew Research Center, U.S. Workers and AI, Feb 2025).

Methodology and Sources

Every statistic on this page was traced to the organization that produced the data: government agencies and statistics offices (NTIA, Eurostat, Statistics Canada, ITU, USPS, FBI), legislatures and regulators, peer-reviewed studies and working papers, company filings, and companies that measured their own platforms or ran their own surveys (Microsoft, Google, Kaspersky, Cloudflare, Litmus, Adobe, Pew Research Center). We do not cite statistics roundups or blogs that repeat other people's numbers. Each link was opened on 2 Oct 2026 to confirm the page loads, the number appears on it and the date is right; PDFs were downloaded and read in full. Telemetry, computer logs, lab tests and surveys are labelled and never combined into one claim, Radicati's figures after 2024 are labelled as forecasts, the Radicati report links to an archived copy because radicati.com did not resolve on 2 Oct 2026, and vendor research is named as such. The page is refreshed monthly, with fast-moving figures such as spam share, email client share and AI adoption replaced first.

Glossary

  • BEC: business email compromise, a fraud in which a criminal poses as an executive, colleague or supplier by email to get a payment or data.
  • DMARC, SPF: email authentication standards that let a receiving server check that a message really comes from the domain it claims.
  • EHR: electronic health record. Its inbox (in-basket) holds patient messages, test results and refill requests.
  • FBI IC3: the FBI's Internet Crime Complaint Center.
  • ITU: International Telecommunication Union, the UN agency for digital technologies.
  • NTIA: US National Telecommunications and Information Administration, which runs the Internet Use Survey.
  • USPS: United States Postal Service.
  • NBER: National Bureau of Economic Research, a US nonprofit that publishes economics working papers.
  • ITT and LATE: intention-to-treat (everyone offered the tool) and local average treatment effect (the people who used it), two ways of reading a randomized trial.
  • Telemetry: usage data measured by software, here anonymized Microsoft 365 signals, as opposed to what people report in surveys.
  • Microsoft 365 Copilot: Microsoft's AI assistant inside Outlook, Word, Teams and other Microsoft 365 apps.
  • Mail Privacy Protection: an Apple Mail feature that preloads tracking pixels, which inflates Apple's share of tracked email opens.
  • Right to disconnect: a legal right or policy that lets employees ignore work contact outside working hours.
  • CHI, CSCW, KDD, WWW, IMWUT, MobileHCI: peer-reviewed computer science conferences and journals where the email and typing studies on this page were published.
  • OBHDP, JAMIA: Organizational Behavior and Human Decision Processes, and the Journal of the American Medical Informatics Association.

Sources

Cite this page

Every figure here links to the study that produced it. Quote any of them with a link to this page or the original source.

WriteVoice. "143 Email Statistics October 2026." WriteVoice, October 2026. https://www.writevoice.io/statistics/email-statistics/
<a href="https://www.writevoice.io/statistics/email-statistics/">Email Statistics</a> (WriteVoice, October 2026)
  • Meeting Statistics Hours spent in meetings, how many are unproductive, what they cost, after-hours calls, follow-up gaps and AI note-taking adoption.
  • Texting Statistics Texts sent per day, who texts most, texting vs calling, voice notes, phone typing errors, texting while driving and text scams.