ManageTime

Article

Ask by email or ask in person: choosing the channel before you send

Ask in person when you need someone to agree to something. Experiments find people say yes far more often to a request made face to face than to the same request sent by email, and the person asking does not expect that gap. Email is the right channel for moving information, not for persuading anyone. This article covers the evidence behind that gap, which asks are worth walking over for, and where email still wins. The research here is not new, and its age is worth knowing before you rearrange how you ask people for things.

Should you ask by email or in person at work?

The message has been drafted twice and it still sits there. You want a colleague to read a document by Thursday, typing it feels like the low-friction option, so you send it and wait. The channel decided that wait. A request worth forty seconds at somebody else's desk now costs you a day of not knowing, and possibly a second message on Friday asking whether the first one arrived.

Sort the message by what you want back, and the decision gets simpler than it looks. If you want information moved, a file, a date, a number, a confirmation of something already settled, email is the better tool and always was. It leaves a written record, it lands when the other person is free to read it, and nobody has to be in the same building. If you want somebody to agree to do something they have not already agreed to, you are sending a different kind of message, and email is the weakest place to send it.

Two behavioural experiments and two supplementary vignette studies, published in 2022 by Roghanizad and Bohns, measured how large that difference is and how well people see it coming. Participants predicted the channel would barely matter, or would not matter at all. When they went out and actually made the requests, the differences between channels were substantial, and the advantage ran to asking face to face over any mediated channel.

So the rule worth carrying: email the things that need recording, walk over for the things that need agreeing. Task Batching is the natural home for the first pile, since a request for information loses nothing by waiting for your next inbox session. The second pile is worth interrupting yourself for. It is also smaller than your inbox makes it look.

Why does asking in person work better than you expect?

Not because your colleagues turn generous in the corridor. The gap sits in your own forecast.

The underlying finding is older than the email question. In 2008, Flynn and Lake ran six studies on what people expect when they ask for help, and found the expectation badly calibrated in one direction. Across the first three, in laboratory settings and in real ones, people underestimated by as much as 50 percent how likely others were to agree to a direct request for help. You predict fewer yeses than you get, whoever you are asking and whatever you are asking for.

The sixth study in that set looked at where the error comes from, and it is not modesty. Someone asking for help thinks about the cost of saying yes: the time, the effort, the inconvenience they are about to hand over. The person being asked is also weighing the cost of saying no, which is awkward, visible, and paid on the spot in front of you. Help seekers were less willing than potential helpers to take that second cost seriously, attending instead to the cost of agreeing. The study did not test channels, so the next step is this site's reading rather than its finding: in writing, the cost of no shrinks to almost nothing. Declining an email takes one line, or no line at all, and nobody watches you do it.

Roghanizad and Bohns tested the email side directly in 2017, in a paper whose title carries the result: you are less persuasive than you think over email. Their 2022 follow-up added the middle of the range, video and phone, and found the same blind spot one step further in. People underestimated the advantage of asking face to face over any mediated channel, and they underestimated the advantage of asking through richer channels, the ones carrying voice and video, over email. The ranking people expected was flat. The real one was not.

Email makes it cheap to say nothing.

Which requests should never go by email?

Start with the ones you have already sent twice. A follow-up on a request that got no answer is the clearest signal available that the channel was wrong, and a third message is usually a way of avoiding a two-minute conversation. Beyond that, three kinds of ask earn the walk. Anything where somebody has to give up time they had not budgeted, so reviewing your draft, covering an hour, joining a project. Anything a reader could take as criticism, since the tone you intended is not the tone that arrives. And anything you need decided today.

Face to face is not always on offer, which is why the 2022 study is useful beyond its headline: it ranked the alternatives too. A phone call or a video call sits above email on the same scale, so when the person you need works in another country, the choice is not email or nothing. It is email or five minutes of somebody's voice. Chat belongs with email rather than with the call, because it is still text arriving without you in the room, and it offers a reader the same quiet exit.

Email deserves a defence here, because a good deal of the advice on this subject reads as though writing were a failure of nerve. It is not.

Write the ask when the answer has to be exact, when the request will travel through three people before it lands, or when either side may want to check the wording months later. Write it when the person you are asking works six timezones away and any live conversation costs one of you an evening. Email is also the kinder channel for a request somebody may want to refuse without an audience. That is the same mechanism read from the other end: the low cost of saying no protects them as reliably as it frustrates you.

Asking in person carries a cost this research does not measure, which is the interruption you cause by walking over. Keep it small. Go when they are already between things, say how long you need before you start, and bring the ask in one sentence so they can answer and get back to work. A remote team has no desk to walk to, and a two-minute call at an agreed hour does the same job there without dragging a meeting invitation behind it. The meetings pillar argues against the half hour that should have been a message. Two minutes of talking to save a two-day thread is the opposite trade, and it usually pays.

How does your working style change the answer?

Which half of this you will actually do depends on how you already handle friction. Readers closest to this site's Architect style prefer the written ask because it is precise and it leaves a record, and their risk runs to three careful emails where one knock would have settled it. Sprinters walk over happily and should watch the other edge. A corridor yes that nobody writes down is not yet an agreement, so send the one-line confirmation afterwards. Improvisers ask whoever is nearest, which is often the right instinct and occasionally the wrong person, so name who owns the decision before you go. Visionary-leaning readers make the case well out loud and lose the particulars, so fix the date and the deliverable before the conversation rather than during it.

None of this transfers cleanly to every workplace, and the evidence is old enough that saying so matters. The newest study here is from 2022, and the requests in this literature were largely made of strangers by student and online participants rather than by colleagues who work together every week. The years since have moved a great deal of office work into chat and video by default. Whether the face-to-face advantage holds its size in a team that has never shared a room is an open question. The direction of the error is the part that has held from the 2008 studies to the 2022 ones: people asking for things expect the channel to matter less than it does.

There is also a case for choosing the weaker channel on purpose. Write when you want a real answer rather than a compliant one, when the person has already refused you once, or when the power gap between you is wide enough that a face-to-face no would cost them something. A yes you got because refusing felt uncomfortable is not information you can plan with. Some asks should stay easy to decline.

Take the message you keep rewriting. If what you want back is a fact, send it. If what you want is a yes, close the draft and find out whether they are at their desk. The email pillar covers what to do with everything still arriving in your inbox, and the working style self-check points at the methods most likely to hold for the way you work.

Frequently asked questions

Should you ask by email or in person at work?

In person when you want somebody to agree to something, email when you want to move information. Experiments published in 2017 and 2022 found people say yes to face-to-face requests far more often than to the same request sent by email, and that the person asking does not see the gap coming. Email keeps its advantages for anything needing a record, exact wording, or a reader in another timezone.

Which work requests should never be sent by email?

Anything asking somebody for time they had not budgeted, anything a reader could take as criticism, and anything you need decided today. A request you have already sent twice belongs in that set by definition. If walking over is not possible, a phone or video call still sits above email on the same ranking, so the choice is rarely email or nothing.

Does asking in person actually get you a better answer?

It gets you more yeses, which is not the same thing as a better answer. The first three of six studies published in 2008 found people underestimated by as much as 50 percent how likely others were to agree to a direct request for help. Part of the reason is that refusing in person is awkward and visible. So face to face is the wrong channel for an ask you want somebody to feel free to turn down.

When is a chat message better than an email?

When the ask is small and the answer is short, and you would otherwise wait hours for a reply that takes ten seconds to type. Chat behaves like fast email rather than like a conversation. It is text, it arrives without you in the room, and declining it costs the reader nothing, so it inherits the same weakness for persuasion. Use it for logistics, not for asking somebody to take on work.

Sources