You reach for your inbox to answer one question: is anything in there on fire. Checking answers it for a minute. Then the question reopens, so you reach again. That’s why willpower does such poor work here. The habit ends when something else starts answering the question for you.
- Three urges get called checking. One of them starts somewhere other than email. Knowing which one is yours decides everything after.
- Checking less has been tested properly, once. Daily stress went down. The amount of mail people got and answered stayed exactly where it was.
- You’ll remember the week as ordinary. The same study asked how stressful that week had been, and the answer held steady. That gap is where most attempts die, somewhere in week two.
- The count of trips beats the tidiness of the schedule. One study measured hearts instead of asking people. Time in the mail client tracked stress, and clustering it into a few sittings came out flat.
- The order matters more than the moves. Arrange to be told first, then cut the trips. Doing it the other way around asks you to carry the uncertainty on willpower, and that’s the version that fails.
- Some of your checking is right. It’s worth separating that out before you change any of it.
This piece is about the reaching, and how to end it. The dread underneath it, and what an inbox does to a person that a task list leaves alone, is why your inbox makes you anxious. What the habit costs in hours and dollars is the real cost of email overload. The four-figure unread count is unread email guilt. The version that follows you home is protecting your evenings. The version with a payroll behind it is email overload when you own the firm.
What you are actually reaching for
Most advice on this treats checking as one behavior with one cause. There are three. They feel identical from the inside, and each one takes a different fix.
The sweep. You want to know that everything’s fine. This is the common one, and it’s the one that fires every few minutes. Underneath it sits a question about risk, and your inbox is the only instrument you have for asking it. The sweep ends when something else can answer that question.
The wait. You’re expecting one specific thing. The signed engagement letter, the reply from the client who went quiet, the confirmation that the wire landed. You’re running a general search several dozen times a day to answer a question with one right answer. That’s an expensive way to ask, and this one has the cleanest fix of the three.
The exit. You’re stuck, or bored, or the next thing on your list is unpleasant. The inbox is the nearest door out of the room. This one looks exactly like the other two and comes from somewhere else entirely. It starts in the work, so the work is where it gets fixed, and people who treat it as an email problem end up deciding the whole subject is theater.
A one-day test sorts them, and it costs you almost nothing. Take one ordinary day. Every time you reach, write one word before you look: sweep, waiting, or stuck. Leave everything else exactly as it is. Most people are surprised by the split, and the surprise usually runs the same way: far more sweeps than they expected, and a stubborn cluster of exits in the two hours after lunch.
For the record, you’re in ordinary company here. The American Psychological Association’s Stress in America 2017 report on technology and social media put the figure at 43 percent of Americans who constantly check their emails, texts or social media accounts. That group reported an average stress level of 5.3, against 4.4 for everyone else. The scale runs 1 to 10, where 1 is little or no stress and 10 is a great deal. Among employed people who constantly check work email on days off, it was 6.0.
Read that carefully before you lean on it. Harris Poll fielded the survey for the APA between August 5 and 31, 2016, among 3,511 US adults, weighted to population proportions. The sample is solid and the vintage is 2016. A “constant checker” there is somebody doing this across email, texts and social media together, and the whole thing is a snapshot rather than an experiment. It tells you how common this is, and the question of what caused what stays open. What it does establish is that this is an ordinary condition, reported by more than four in ten adults. That’s worth knowing if some part of you had filed it under personal failing.
Checking less works, and you will not remember it working
One experiment took this question seriously, and its most useful finding is the one that gets quoted least.
In 2015, two researchers in Vancouver, Kostadin Kushlev and Elizabeth Dunn of the University of British Columbia, published Checking email less frequently reduces stress in Computers in Human Behavior. They gave 124 adults two weeks. One week meant checking email three times a day, with the mailbox shut and the alerts off. The other meant checking as often as possible, with everything open and pinging. Everyone did both, in randomized order. The design, and what it proves about volume, is handled in triaged the moment it lands, the page next door to this one on the mechanics.
The headline results came out the way you’d hope. Daily stress was significantly lower in the limited week, reported as F(1, 121) = 4.18, p = .04, with a Cohen’s d of .37. While doing something that mattered to them, people felt calmer, F(1, 96) = 3.84, p = .05, d = .45. Modest effects, and real ones.
Now the finding that changes what you do on a Tuesday. At the end of each week the same people were asked how stressful that week had been overall. The answer held completely steady: F(1, 91) = .04, p = .838, d = .04. The authors put it plainly, that limiting the frequency of checking “did not affect their memory of how stressful the week had been overall.”
Sit with that for a second, because it’s the whole practical problem in one line. The days went better. Memory of the week stayed flat. The benefit was real, and the person receiving it lost sight of it the moment they looked back.
That’s exactly where attempts fall over. You try checking less. Around Thursday of week two you take stock, ask yourself honestly whether you feel better, and get an honest shrug. Then you file the advice under productivity theater. The study says you’d give that answer even while it was working. So audit this by counting. The number of times you opened the mail client is a fact, and it’s the one signal visible to you at this range.
One more honest note from the same paper, because it narrows who this is for. The drop in feeling distracted by email held for the students, who were about two thirds of the sample. For the community members it faded, and their post-hoc comparison came out at p = .89. The authors say outright that they recruited heavy email users with some flexibility in their day, so the intervention “might be unlikely to reduce stress among individuals who receive little email or have no choice about how frequently they check email.” If your job is dispatch, support triage, or anything where the mail is the work, the rest of this piece is about a different job.
What batching does and does not buy you
The standard prescription out of that study is batching. Gather your email into two or three deliberate sittings, and leave it alone in between. Somebody measured that in the field rather than handing it out as an instruction, and what came back is worth having before you build your week around it.
At CHI 2016, Gloria Mark of the University of California, Irvine, with Shamsi Iqbal, Mary Czerwinski and Paul Johns of Microsoft Research and Akane Sano of the MIT Media Lab, published Email Duration, Batching and Self-interruption. Forty information workers in a research division of a large corporation wore chest heart rate monitors during all waking hours. Their computer activity was logged alongside it, for an average of 12 workdays. Stress came from heart rate variability, measured continuously rather than asked about.
Three results, and they point in different directions.
Time on email tracked stress. The more of an hour someone spent in the mail client, the higher their measured stress in that hour. They also rated their own productivity lower at the end of the day. That relationship held after controlling for job characteristics.
Batching came out flat. The researchers sorted people into batchers, consistent checkers and a mixed group, by looking at how each person’s email time was spread across the day. That gave them 11, 23 and 6. Batching “was neither significant as a main effect on Stress, nor was there a significant interaction of Batching behavior and Email Duration on Stress.” Their own summary is blunter: “contrary to claims that batching email can reduce stress, we found no evidence in our study to suggest this.”
The source of the interruption came out flat too. Some people checked mainly because a notification told them to, and some mainly because they decided to. That difference showed no significant relationship with stress. If anything the self-directed checkers came out slightly ahead on their own sense of productivity at high email volume. The authors read that as agency: choosing your own moment lets you break at a seam in the work rather than mid-sentence.
Take those with their limits on. Eleven batchers in one corporation, observed rather than assigned, is a long way from proof. Heart rate variability and self-reported stress are two different quantities, so this study and the Vancouver one can both stand. The combination is still instructive. Checking less lowered how stressed people said they were. Arranging that same checking into tidy clusters left their hearts where they were.
The reading worth acting on: aim at total trips and total time. Two long sittings with phone sweeps in between leave you where you started. Three fewer trips a day counts, even when they land at ugly times.
The order the changes have to go in
The order below does real work. Almost every failed attempt starts around step four, which is the point where you’re carrying the uncertainty on willpower alone.
1. Make it possible to be told. Before you reduce anything, arrange for the one thing you’re afraid of to reach you anyway. A specific named person. A subject line you’re waiting on. A word that shows up only in the message that would ruin your week. If your mail app can raise an alert on a single sender or a single phrase, that’s the feature you want, and one rule beats twenty. Then test it once on purpose. Have somebody send the thing, and watch it arrive loudly. You’ll believe it once you’ve seen it work, and the sweeping eases from there.
2. Answer the waits individually. For every specific thing you’re waiting on, write it down with the date you’ll chase it. A dated item, rather than a flag or a starred message. Something you’ve agreed to look at on Thursday stops asking to be looked at today, and the relief runs well ahead of the effort. This is the highest-yield hour in the whole list, because the waits are what make the sweeps feel justified. Keeping track without a task list is the mechanics of it.
3. Remove the manufactured checks. Badge counts, banner previews, lock-screen text and the little envelope in the corner each invent a reach of their own. The badge is the worst of them. A number with nothing behind it is pure uncertainty with a red circle around it. Turn off the count as well as the sound. Move mail off your phone’s first screen, so getting there takes a deliberate act. This step does part of the job, and the CHI 2016 result says which part. Silencing alerts turns interruptions somebody else caused into interruptions you cause. That swap left measured stress where it was. The manufactured reaches go. The reason stays.
4. Close it. Closed, rather than minimized or parked in another tab. In the study that produced the result, the limited week ran with the mailbox shut and the alerts off, both. An open mail window with a live count in it reissues its invitation every ninety seconds. Intention loses to anything sitting in your peripheral vision.
5. Set the interval from what you owe people. Your real response window is probably still unwritten, so write it down now. How long would a client, a colleague or a supplier wait for a reply from you and think it perfectly normal? For most professional work the honest answer is hours. Build the schedule to fit that, with margin. Then run one test on it: would you be willing to publish it? Picture telling a client this. “I read email at 8:30, 1:00 and 4:30, and my number is on the signature if it cannot wait.” If saying that out loud makes you wince, what you’ve set is a fear with a schedule painted over it. Three a day is a research protocol. Yours is whatever your obligation supports.
6. Tell two or three people. The authors of the Vancouver study flag their own biggest limitation right here. They changed what participants did, and they “had no control over the expectations of those around them,” so they suggest a team-wide or company-wide version would do more. A company-wide version is out of your hands, and one sentence to your three highest-volume correspondents is well within them. Add one line to your signature. Give people a faster channel at the same time. The aim is to stay reachable in a way that leaves you free to look away.
7. Give the exits somewhere else to go. If your one-day test turned up a cluster of stuck reaches, treat that as a work problem. The reach at 2:40 in the afternoon tells you the current task is missing its next physical action. Deciding what that action is takes about thirty seconds, and it takes away the reason to leave the room.
8. Count, and only count. Once a week, write down the number of separate visits. That’s your instrument. Feelings stay flat here for a good while after the change starts working, so let the count do the judging.
Two things sit off that list on purpose. Getting the count to zero resets a number, and the reaching was always about a different question, which a calm inbox versus inbox zero works through. And resolving to try harder puts the blame on the one part of you that’s been working correctly all along.
How long this actually takes
Longer than you’ve been told, and the honest number comes from a study of the opposite process.
In 2010, Phillippa Lally and colleagues at University College London, in the United Kingdom, published How are habits formed in the European Journal of Social Psychology. Ninety-six volunteers repeated one chosen daily action in a consistent setting for 12 weeks, rating how automatic it felt as they went. Among the 39 whose data fit the model well, “the median time to reach 95% of asymptote was 66 days, with a range from 18 to 254 days.”
Two caveats, both real. That’s a British sample of mostly postgraduate students doing things like drinking a glass of water after breakfast. And it measures a habit being built rather than one being taken apart, which is its own separate process. The equivalent study for email is still waiting to be run. What the number is good for is calibration. Automaticity arrives in months, and the spread between people is enormous. If you’re on day 20 and it still takes effort, you’re on schedule.
The same paper carries the most reassuring finding in this whole piece, and it’s the one to hold on the day you break the rule. Missing a single opportunity left the process intact. Their conclusion is that “missing one opportunity does not preclude habit formation, but missing a week’s worth of opportunities reduces the likelihood of future performance and hinders habit acquisition.” One bad Tuesday is one bad Tuesday. Two bad weeks is a different thing. So the recovery move after a slip is your next scheduled check, today, rather than a fresh start on Monday.
And the ending is quieter than you’re expecting. The habit goes unpaid, day after day, and eventually stops turning up. The quiet confidence of never missing mail is that written out as an ordinary evening, which is how it actually arrives.
When the checking is the right call
Three cases where the advice above steps aside, and it’s worth being straight about all three.
Your job is the mail. Support queues, dispatch, front-desk coverage, a shared inbox during filing season. The Vancouver authors scope their own finding away from exactly this population. If the messages are the work rather than the interruption to it, checking constantly is the job description. What helps there is a way to share the load, rather than a way to look less.
You are genuinely on the hook today. A deal closing, a patient, a family situation, a client mid-crisis. Suspend the schedule out loud, rather than letting it quietly lapse. An interval you break on purpose is one you can pick back up afterward.
Something else is coming to rest here. If the pull is there on days you’re away from email entirely, or shows up well outside work, or has been getting heavier over months, then your inbox is where something else is landing. Why your inbox makes you anxious draws that line properly, including the point where the useful next step is a doctor rather than a better setup. A calmer inbox is a good thing to own, and it sits alongside treatment rather than in place of it. Both are worth having at once.
What changes when you can be told instead
Everything above turns on one substitution. The sweep stops when something else starts answering the question you’ve been asking by hand.
That’s the job Point does. You describe the thing you’re waiting for in one sentence. Point watches everything coming in and reaches you once, at the moment it matters. That’s the direct replacement for the sweep.
For the waits, a request arriving inside a thread becomes an item with a date on it, so you get to put it down. Open loops come back to you while they’re still warm, which is nothing falls through the cracks in full.
And when you do sit down, you open a short ranked feed. Every thread carries a plain summary, so a long conversation lands in a few lines and the trip is genuinely shorter. Time in the mail client is the quantity the CHI 2016 study found tracked stress. You also choose when you’re brought up to speed, and what’s worth an interruption in between. That’s the interval decision, made once rather than forty times a day.
How much Point handles unaided is yours to set, one kind of action at a time. Each kind sits somewhere on a dial: suggest-only at one end, review in the middle, fully handled at the far end. Out of the box, each one sits on review. Point prepares the work and holds it for you. Move a single kind of action up to the top of the dial, and Point handles that one without asking. Point’s own actions are logged and reversible, with the one exception every piece of software shares: a message already delivered sits on somebody else’s server, out of reach. The whole shape of the week is on the benefits page, and being told the moment it happens is the part aimed squarely at this habit.
Here’s the honest edge of it, because this is a piece about a compulsion and overselling would be its own small unkindness. The same volume of mail still arrives. Your real response obligation is still yours to decide, which is step five. The months the habit takes to starve run their full length either way. And if your one-day test came back full of exits rather than sweeps, the work is where the repair goes. What Point takes away is the reason to look, and that’s the one lever in this whole piece running on something other than willpower.
Common questions
How many times a day should I check my email?
The number that works is the one you derive yourself. Three a day was a research protocol, chosen so a mixed group of volunteers could plausibly comply with it, and the authors offered it as a study condition. Start from your real response obligation, then apply the publication test. Pick a schedule you’d be willing to tell a client about in writing, with a faster channel offered alongside it. If stating it out loud would embarrass you, it’s a fear wearing a decision’s clothes, and it’ll come apart.
Does turning off notifications stop the checking?
It stops one kind of it. Silencing alerts clears the reaches somebody else caused and leaves the ones you cause. In the CHI 2016 field study, the split between people who checked mainly on notifications and people who checked mainly on their own showed no significant relationship with measured stress. Do it anyway, especially the badge count, because a number with nothing behind it is pure uncertainty. Expect it to be step three of eight, with the other seven still to do.
Is batching my email into two or three sessions the fix?
It’s one part of it, and the evidence behind it is thinner than its reputation. When batching was measured in the field with heart monitors rather than handed out as an instruction, it showed no effect on stress at all, while total time on email showed a clear one. Fewer trips and less time in there are the quantities that move. Arranging the trips into elegant blocks looks like a matter of taste.
I only check because I am waiting on one particular reply. What do I do about that?
That one has the cleanest fix in the piece. You’re asking a specific question with a general search. Set an alert for that one sender or that one phrase. Put a dated item on the calendar for the day you’ll chase it if it’s still quiet. Those two moves together retire the whole category, and for most people the waits are what make the general sweeps feel defensible in the first place.
How will I know it is working?
By counting. This is the least intuitive finding in the research and the most useful. The same experiment that lowered people’s daily stress left their sense of the week exactly where it was. Looking back on the week is a poor instrument here. The number of separate visits per week is a good one.
Is checking constantly a sign of something wrong with me?
It’s an ordinary response to uncertainty, and on the APA’s own figures more than four in ten American adults describe themselves as doing it. The useful test is whether it stays inside email. If the same pull is there on days you’re away from your inbox, or has been getting heavier over months, a doctor is a better first stop than a new email setup.
The short version
You check because you have a question, is anything in there on fire, and one instrument for asking it. Checking less genuinely helps. The one experiment on it found lower daily stress, with the same amount of mail arriving and the same amount answered. That same experiment found memory of the week holding steady, which is exactly why most attempts get abandoned in the second week. The benefit is real and invisible at the same time.
So do it in the right order. Arrange to be told about the thing you’re afraid of, and test that once. Give every specific wait a date. Strip the badges and previews that manufacture reaches. Close the window rather than minimizing it. Set your interval from what you actually owe people, then say it out loud to the two or three people who write to you most. Judge it by a count of visits, and let the feeling of the week go. Expect months rather than weeks, and treat one slip as one slip. And if the reaching is really about the work you’re avoiding, or shows up on days you’re away from email entirely, the repair belongs somewhere else, and that’s worth knowing early.