Snoozing an email takes it out of your inbox and brings it back at a time you choose. Setting mail aside is the same move under a plainer name. It is the second half of every triage decision, and the half an ordinary inbox records worst: it offers you several ways to mark that a message is not for now, and nowhere at all to put the when.
- Deferral is not the exception in a triage pass. Most of what a pass produces is mail you have decided not to deal with yet.
- A flag records that something exists. It does not record what the message wants or when you want it back, and people mostly do not find their way back through it.
- A snooze answers one kind of later, a time. The laters that are really events, or really conditions, need something else.
- Out of sight is the whole point and the whole risk. Mail you cannot see is mail you cannot count.
What does snoozing an email do?
Three things, in order. The message leaves the list you are looking at. It waits somewhere it can still be found. And at the hour you named it comes back to the top, behaving like new mail.
Only the third is unusual. Everything else an inbox offers you does one or two of those and stops short of the last one, which is the one that matters.
Archiving does the first two. The message leaves, it stays in your mail, and nothing brings it back, so finding it again depends on you remembering it exists. A flag or a star does neither of the first two: the message stays exactly where it was, at whatever depth the day pushes it to, wearing a small mark. Marking a message unread is the strangest of the three, because it works by throwing away a fact. The one thing you knew about that message was that you had read it, and the price of the reminder is losing that.
Both mainstream clients have the feature now, under the same name. Google’s help says that when you snooze, “the email is removed from your inbox temporarily”, and that it “comes back to the top of your inbox when you want it to, whether that’s tomorrow, next week, or in the evening”, waiting under a Snoozed label until then. Microsoft’s description of the Outlook version is shorter and the same shape: a snoozed message reappears “at a specified time” (both checked 18 August 2026). The mechanism is neither exotic nor new. What is interesting is not whether you have it, but what it can and cannot express.
One more property is easy to miss. A snooze is entirely private. The sender’s copy is unchanged, their thread still shows no reply, and nothing about the arrangement reaches them. A snooze buys you quiet and buys the person waiting on you nothing at all, which is worth holding on to when the thing being deferred is somebody else’s decision.
Why most of triage ends in later
Triage ends in three answers and only one of them is now. How email triage works sets out the pass and where it breaks under volume. What concerns us here is the pile it leaves behind.
The closest measurement of that pile is Characterizing and Predicting Email Deferral Behavior, a 2019 study for the ACM conference on web search and data mining by Bahareh Sarrafzadeh, Ahmed Hassan Awadallah, Christopher H. Lin, Chia-Jung Lee, Milad Shokouhi and Susan Dumais. It runs on two tracks: fifteen interviews with Microsoft staff, conducted in their own offices with their own mailboxes open, and a log analysis of forty thousand enterprise users of Outlook’s web client across the fortnight from 6 to 19 May 2018, some three million recorded actions. Their definition is the one worth keeping. An email is deferred when “the user deliberately postpones completing the task related to it to later”.
All fifteen said they defer, routinely. One put the rate at half of everything that arrives.
The logs read lower, and the gap is instructive rather than contradictory. Measured through actions, 16% of active users deferred at least one message a day on weekdays, deferral showed up in at least 12% of triage sessions, and 3% of all messages were deferred. Those are floors, because the only deferral a log can see is one that ends in a reply, a reply-all or a forward. Something you set aside and then settled by phone, or read properly and closed, or simply never got to, is invisible to the count. A second figure gives the better sense of shape: around 10% of messages get a reply, reply-all or forward at all, and 26% of those are written in a later session than the one where the message was first read. A quarter of your replies are already being deferred, and nothing in a mail list is built for that.
What gets deferred is not what you would guess. The interviews produced five factors, ranked by how often they came up: how much time and effort handling it would take, who sent it, how many other people were on the thread, the reader’s own workload at that moment, and last of the five, urgency. The message you put down is usually the one that would cost you twenty minutes rather than the one shouting, which is a separate observation from the one important versus urgent makes and points the same way.
The logs agree from the other side. Deferred messages averaged 3.9 recipients against 7.0 for messages handled on the spot, and carried roughly twice the rate of explicit requests and reply requests. They were likelier to come from a person than a machine, likelier to come from an important sender, and, quietly, less likely to come from a sender the reader had heard from before. Addressed to you, written by a human, asking for something, from someone you have no history with. That is also a fair description of a new client.
It gets worse exactly when you can least afford it. The same logs show deferral rising with workload: the more unhandled mail already sitting there, the likelier the next message was to be put down, and the probability climbed from about 3% to about 4% as the meetings on a calendar went from one to five. The deferred pile grows fastest on the days with the least room to service it.
Fifteen people at one software company and forty thousand enterprise Outlook accounts are not a five-person practice, so take the direction rather than the decimal. The direction is not in doubt.
Why a flag is not a return date
Here is the awkward part. People do mark the mail they intend to come back to, and the marks turn out not to be what brings them back.
A companion study measured it directly. The Lifetime of Email Messages, presented at the 2018 conference on human information interaction and retrieval by Tarfah Alrashed, Ahmed Hassan Awadallah and Susan Dumais, paired a log analysis of one hundred thousand active enterprise users of Outlook’s web client, February to May 2017 and some 800 million actions, with a survey of 395 Microsoft employees in the United States, each asked to recall the last message they had gone back to.
Three of its numbers sit badly together.
People mostly saw it coming: 62.3% said they had known at first reading that they would need to return to that message. Then, of everyone surveyed, 38.5% did nothing at all to make the return easier, 27.4% moved the message to a folder, and 19.2% flagged it or marked it unread. And when the moment came, 71.6% found their way back by searching for it. Browsing the inbox or a folder accounted for 20.9%. Filtering for flagged or unread mail accounted for 1.5%.
One and a half percent. The marker is not the road back, even for the people who set it.
The deferral study reaches the same place from the other end. Its authors pulled every message in their sample that had been flagged or marked unread, 10,551 of them, and asked whether each went on to get a reply, a reply-all or a forward in some session after the one it was read in. About 14% did. Their own summary of that result is the sentence to keep: it suggests “users often do not go back to messages they explicitly mark”.
Two caveats belong with it. The window was a fortnight, and a message can be settled by a phone call or a fresh thread rather than a reply, which their interviews describe people doing. So 14% is not the share of marked mail that was neglected. It is the share that came back as a reply, which is what people said they had in mind when they marked it.
The reason is not mysterious. A flag carries one bit: something here. It cannot say what the message wants, so you reread it to find out, and it cannot say when, so the choosing happens again on every pass. It does not even carry the same one bit for everybody. The interviews found flags meaning “still pending” to people who let mail pile up, and meaning “important or urgent” to people who keep the inbox or the unread count at zero. A signal that means two things carries less than one.
Filing the pile away instead has its own failure and its own guide, because a deferred queue sorted into action and waiting and someday is a ranking hidden from yourself rather than a ranking done: inbox zero versus a calm inbox.
A snooze fixes precisely one thing here, and it is the right thing. It attaches a when, so the message stops depending on you noticing it.
What kind of later is it?
One of the people interviewed in the deferral study said the useful thing outright: “every time you flag something for later you should indicate what kind of later it is.”
There are three kinds, and only one of them is a time.
A later that is a time. Thursday morning. After the quarter closes. When I am back from leave. This is what a snooze expresses, and it expresses it exactly.
A later that is an event. When the client returns the signed copy. When the bank confirms. When the surveyor’s report lands. No hour presents itself here, and naming one is a guess in both directions: too early and you look at it again for nothing, too late and you found out late. Most of what people describe as waiting is this kind, and a snooze is the wrong instrument for it.
A later that is a condition of your own. When I have half an hour and the file open. Another participant in the same study: “when I have 30 minutes I want to devote it to something longer […] and then for easy tasks I’m gonna do it when I have five minutes and I won’t do it when I have 30 minutes.” The researchers found people choosing the moment by optimizing two things at once, how little context switching it would cost and how few separate sittings the message would take. That is a genuine scheduling problem, and you are being asked to solve it mid-pass, in a second or two, about a message you have only skimmed.
Which is the honest limit of a snooze button. It answers the first kind well and hands you the hardest part of the job: choosing an hour, one message at a time, at the worst possible moment for choosing anything. That work does not vanish because the button exists. It vanishes when something else can propose the time, or watch for the event rather than making you name an hour that stands in for it.
And when the later has a date in it because the message asked for something by then, it was never a deferral. It is an obligation with a deadline, which wants a different mechanism and a better one.
Where snoozing goes wrong
Four ways, and three of them are one mistake in different clothes.
The message snoozed twice. A return time you cannot act on produces another guess, made with no better information than the first. Repeated snoozing is rarely a discipline problem. It is usually a later of the second kind dressed as one of the first: you are waiting on an event and expressing it as a date, so the date keeps arriving before the event does.
Monday morning. Snooze menus offer round times, and round times collect. Tomorrow, next week, Monday morning. What returns is not one message at a considered hour but a small avalanche, landing on top of everything that arrived overnight of its own accord, at the hour of the week you can least absorb it. Deferral moved the pile. It did not shrink it, and it can stack it.
The count you gave up. Out of sight is the feature and also the cost, because you can no longer tell at a glance how much you owe. If what you have set aside cannot be seen as one list with its return times on it, you will start keeping a copy in your head, which is the job you were trying to put down.
Snoozing instead of declining. Some mail is deferred because you would rather not answer it, and a snooze is an excellent way to keep not answering it, on a schedule, with a clear conscience. The tell is the third return of the same thread. What that message is usually waiting on is a decision, and the decision is usually no, or a two-line answer you have been treating as a longer job than it is.
How Point sets mail aside
Start with the part that changes the arithmetic. In Point everything that arrived overnight is read and ranked before you open the app, into now, later and other, with how much a message matters and how soon it is due weighed as two separate things. The not-now call has already been taken on the bulk of the pile, so what you set aside by hand is the exception rather than every message, and choosing an hour becomes affordable.
When you do set something aside, you pick the moment it comes back and it returns on its own at that moment, into the same ranked feed. Later is a section of that feed rather than a hole you have to remember to look into.
The second kind of later has its own mechanism, and this is the part no snooze button reaches. Point keeps track of what you are waiting on and reminds you to chase it at the point chasing is worth it, and a thread that has gone quiet comes back to you before it goes cold rather than at an hour you guessed at three weeks earlier. You can also ask in plain words to be told when a particular message arrives or a condition is met, and that request sits there watching what comes in. That is an event expressed as an event.
Where a message asked something of you with a date attached, Point lifts the ask out as a task carrying its due date and a link back to the sentence that produced it, so it never becomes a snooze you have to interpret later. Turning an email into a task covers that properly.
The controls sit around all of it. How far Point goes on its own is a separate setting per kind of action, running from suggest-only through review to handled outright, and review is the default out of the box. What it does is written down in order in plain language, and a row in that record is one click from being taken back. The full inventory is on the benefits page.
Common questions
Where do snoozed emails go?
Not away. A snoozed message is held out of the inbox view and stays findable while it waits, then returns to the top at the time you set. Google describes Gmail’s version as the email being “removed from your inbox temporarily”, waiting under a Snoozed label; Microsoft describes Outlook’s as a message that reappears “at a specified time” (both checked 18 August 2026). It is not a deletion, and it is not a reply.
What is the difference between snoozing and archiving an email?
Archiving decides the message is finished with. It leaves the inbox, stays in your mail, and nothing brings it back, so finding it again depends on you remembering it exists. Snoozing decides the message is not finished with and not for now, so it leaves and returns. Reaching for archive where you meant snooze is the common way something disappears entirely.
Is snoozing better than flagging an email?
For deferring, yes, and the difference is the date. A flag records that a message matters without recording when you want it, so the choosing has to be redone every time you scan the flagged list. In the survey above, filters for flagged or unread mail accounted for 1.5% of how people actually got back to a message. A snooze is the only common marker that brings the message to you instead of waiting to be looked for.
Why do I keep snoozing the same email?
Usually because what you are waiting for is not a date. If the message really depends on someone else replying or a document arriving, every hour you pick is a guess and will keep arriving before the thing does. That case wants a tracked follow-up or a standing watch rather than another snooze. The other common reason is that the message is waiting on a decision you have not made, in which case the snooze postpones the decision rather than the email.
Should I snooze it or turn it into a task?
Snooze it when the message is fine as it stands and simply belongs at a different hour. Make it a task when it carries an obligation with its own deadline, because a task records what is owed and by when, where a snooze records only when to look again. The rule of thumb: if you would have to reread the message to remember what it wanted, it should have been a task.
The short version
Deferral is most of what a triage pass produces, and the tools email gave it are a flag, an unread mark and a folder, none of which record when you want the thing back. The measurements are unkind about how well that works. People know at first reading that they will return, mostly do nothing about it, and when they have marked a message they overwhelmingly find their way back by searching instead. A snooze is the one common repair, and it repairs exactly one kind of later, a time you can name. The laters that are really events want something watching for the event, and the laters that are really obligations want a dated task. What makes any of it affordable is not having to make the call on every message, which is the ranking being done before you sit down.
What ends up in the deferred pile depends on the work. It reads differently in an accounting practice than in a law firm, and there are eight versions of the morning written out.