Key takeaways
- An interruption costs about 23 minutes of refocus; an open inbox produces dozens a day.
- Limiting checks to three a day lowered stress in a randomised trial with no loss of output.
- In a 40-person field study, batching went with higher rated productivity, but stress tracked total email time, not the checking pattern.
- Batching only holds if each batch is small; do the drafting and chasing before the check.
The tab is always there. Every new message is a small decision: read it now, or not. The research on interruption says those small decisions are not small at all, and the research on batching says the obvious fix works, with one caveat most articles leave out.
How long does it take to refocus after an interruption?
Gloria Mark's field studies at UC Irvine followed knowledge workers through their days and found that after an interruption it took, on average, about 23 minutes to return to the original task at the same level of focus. People compensated by working faster, which raised stress and error rates. An inbox that pings is a device for producing that interruption dozens of times a day.
The cost is not the seconds spent glancing at the notification. It is the resumption: finding where you were, reloading the context, and getting back to the depth you had. For anyone whose work needs a held state in the head, the arithmetic is brutal, and the developer version of it is in interruptions cost developers 23 minutes.
Does checking email less often reduce stress?
The experimental evidence is cleaner than the field evidence. Kushlev and Dunn's crossover trial had 124 adults limit checking to three times a day for one week and check freely the other. Daily stress was significantly lower in the limited week, with no measurable loss in getting things done. The participants found it hard to resist checking, and resisting is what produced the benefit.
PeakLook does this from your own data. Drafts the reply. Keeps the follow-up going. Join the free beta →
Does batching email make you more productive?
Mostly yes, with a caveat. Mark, Iqbal, Czerwinski, Johns and Sano logged 40 information workers for 12 workdays with computer logging, biosensors and daily surveys. The longer people spent on email each day, the lower their perceived productivity and the higher their measured stress. Batching was associated with higher rated productivity. But the authors found no evidence that batching email lowered stress, despite the widespread claim that it does.
Read together, the two studies say: fewer checks lowered stress when it was imposed for a week; in the wild, batching improved how productive people felt, and the stress tracked total email time rather than the checking pattern. The lever that moves both is spending less total time in the inbox, which batching helps with only if each batch is short.
| Study | Design | What changed | What did not |
|---|---|---|---|
| Mark, Gudith, Klocke (2008) | Lab task with interruptions | Faster work, higher stress and workload under interruption | Task completion |
| Kushlev & Dunn (2015) | 124 adults, randomised crossover, one week each | Lower daily stress with three checks a day | Getting things done |
| Mark et al. (2016) | 40 workers, 12 workdays, logging and biosensors | Higher rated productivity with batching; stress rose with email duration | Stress, from batching alone |
So how often should you check email?
Three times a day is the only number with a randomised trial behind it. Two to four fits most roles. What matters more than the count is that each check is a scheduled task with an end, not a tab left open.
| Role | Checks a day | Why |
|---|---|---|
| Maker (engineering, writing, analysis) | 2 | Long focus blocks are the output; mid-morning and late afternoon |
| Manager or founder | 3 | The Kushlev and Dunn schedule; people depend on replies within the day |
| Sales, recruiting, support | 4 | Replies are the job, but a schedule still beats a feed |
Turn notifications off between the checks. Genuinely urgent things arrive by phone, and a three-hour delay on everything else is invisible to the sender.
Two further rules make the schedule hold. Put the sessions in the calendar as appointments, so a meeting cannot be booked over them and a colleague can see that you will answer at two. And end each session with the inbox decided, not merely read; a session that leaves twenty threads open just moves the interruption into your head.
Why batching fails by Wednesday
Batching fails when the batch is too big to finish. If three checks a day mean three sessions of forty unread threads, each needing a reply written from scratch, the sessions overrun and the discipline breaks by Wednesday. Batching works when each session is mostly reading and deciding, because the drafting and the chasing have been done before you opened the tab. That is also why the Mark study found no stress benefit from batching on its own: a long batch is still a long time in email.
How to make each batch small
- Filter the noise before it lands. Newsletters, receipts and notifications skip the inbox and go to a folder you read weekly. The inbox zero method covers the filters for Gmail and Outlook.
- Have drafts waiting. A reply you edit takes a quarter of the time of a reply you write. Whether that draft comes from a template or an assistant, it should exist before the batch starts.
- Move chasing out of the inbox. Threads waiting on someone else belong in a schedule, not in your attention. A follow-up sequence sends on day 3, 7 and 14 without you.
- End the session on a decision for every thread. Replied, scheduled, delegated, or deliberately ignored. Nothing stays open.
Which of those jobs a machine should do, and which it should not, is the subject of what an AI email assistant can and cannot do. The size of the whole problem, in hours a week, is in how much time we spend on email.
How PeakLook shrinks the batch
That is the job PeakLook does between your sessions: scan the Focused inbox in Microsoft 365, skip the newsletters, draft a reply for every thread that needs one, and run the follow-ups that are due. When you open the inbox, the drafts are waiting and the chasing has happened. The batch shrinks to the decisions, and three times a day becomes enough.
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.
Questions
Is it bad to check email constantly?
What is the best time of day to check email?
What about urgent email?
Does turning off notifications help?
Will people notice I reply in batches?
Sources
- Mark G., Gudith D., Klocke U. (2008). The cost of interrupted work: more speed and stress. CHI. https://www.ics.uci.edu/~gmark/chi08-mark.pdf
- Kushlev K., Dunn E.W. (2015). Checking email less frequently reduces stress. Comput Human Behav. https://www.sciencedirect.com/science/article/abs/pii/S0747563214005810
- 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/
Builds the UPeak apps and writes up the research behind them.