How to follow up without nagging

A bare bump carries no information and adds only guilt. What a second message has to contain to be worth sending, and why the count matters more than the interval.

6 min read

"Bumping this." "Just circling back." "Any update on the below?"

Read those the way the recipient does. Each reports a fact they already know — you are still waiting — and invites them to feel worse about it. That is the entire payload. Nagging is not a matter of frequency or tone; it is what happens when a follow-up carries nothing except the fact that you sent it.

So the question is not how long to wait or how gently to phrase it. It is what the second message contains that the first one didn't.

A follow-up needs a reason that isn't your impatience

Three things legitimately produce a second message.

A date arrived. The deadline you named is now close, or past. If you named a default — what you would do without an answer — the follow-up is the courtesy notice that you are about to do it.

Something changed. The second quote came in, or the client moved the date, or you found the number they would otherwise have had to ask you for. New information is the only kind of follow-up that is more useful to the reader than to you.

You narrowed the ask. The first message asked for a review of the whole plan; the second asks whether one number is wrong. Different question, and a much smaller one.

If none of the three is true, you have an impulse rather than a follow-up. Wait longer, drop it, or get the answer elsewhere — and if you get it elsewhere, say so, because that is new information too.

Restate the ask in full, every time

Assume they never read the first one. Not as a rebuke — as the base case. Non-reply usually means the message was triaged rather than rejected, and something triaged was never read past its first two lines.

Which makes "see below" an expensive request. Repeating two hundred words costs a reader who did see the original about ten seconds. Sending a reader who didn't back through the thread costs minutes, and usually produces nothing.

Write the second message to stand alone — what you want, from whom, by when, plus the two or three facts needed to answer it. In a RelayLink briefing that is structural rather than something you have to remember: a reply carries its own ask, TL;DR and context brief like any other package. The ask field is capped at 400 characters, which is about one question a busy person can answer and far too short for a recap of the thread.

Make the second one easier to answer than the first

A follow-up that repeats the original question offers the reader the same work they already declined once. Change the work.

  • Offer a default. "Unless I hear otherwise by Thursday, I'll go with B." Silence becomes a legitimate answer rather than a failure, which is why it stops reading as pressure.
  • Convert it to a yes or no. "Do you want me to go with B?" is answerable in a lift. "What do you think?" is not.
  • Cut it to the blocking piece. Ask for the one number or approval you cannot proceed without, and explicitly release them from the rest.

If your message carries a stated response shape, this is where it changes. RelayLink's five are opinion, decision, review, info and fyi, and moving from review to decision is a real edit to what you are asking for, not a relabel.

Decide the count before you send the first one

Pick a number in advance — two is usually right — and honour it. The number matters less than the fact that one exists, because a sequence with no visible end is indistinguishable from an automated one, and readers treat it accordingly. Where it helps, say it out loud. "I'll ask once more next week, then leave it with you" turns an open-ended obligation into a closed one, and closed obligations get answered more often.

The count is also checkable rather than remembered. list_threads returns a package count per thread, so how many times you have written into a conversation is a number you can read instead of a feeling you have. thread_status sorts threads into ones awaiting you, ones awaiting others and idle ones, and for the awaiting-others group it reports either "they have seen it" or "not seen yet" for the most recent package.

Read that narrowly. Seen means their assistant pulled the package or someone opened the web view. The notification mail is plain text with no tracking pixel, so a person who read your note and TL;DR in the email without clicking through still shows as not seen. It is evidence in one direction only.

Give them an exit that isn't silence

Most non-replies are not refusals. They are people who cannot say no cheaply. So make it cheap — "if this isn't yours, tell me who to ask", or "if you'd rather not, just say so and I'll stop." Saying no in writing is a permission most people have never been given, and you can grant it in one sentence.

There is a structural version of the exit, and it is not yours to grant. Every RelayLink notification carries a one-click unsubscribe that blocks the sender for good, with no account and no form. Offering a graceful exit in prose is largely how you keep the exit from being that one.

Free to produce is why the discipline matters

An assistant will write the follow-up in seconds, in the right register, with the ask correctly restated. Every rule above just got cheap — which also removes the one thing that used to limit follow-ups, namely that writing them was tedious. The scarce resource was never the writing. It was the reader's attention, and faster drafting did nothing to that.

RelayLink will not pace you either. There is no scheduled send, no reminder and no snooze, and following up inside an existing thread is never blocked by the daily cap on reaching new people. The interval and the count live in you. What the product imposes is deliberation rather than restraint — an unconfirmed draft can no longer be sent after 24 hours, and trying cancels it and tells you to draft again. Sleep on a follow-up for a day and you have to make it a second time, which taxes exactly the ones that were really impatience.

The test for any follow-up is whether the reader would be worse off not receiving it. If the honest answer is no, the message is for you rather than for them. When several have gone out and nothing has come back, reading that silence is the next problem. To put the ask, the shape and the urgency into fields rather than tone, connect your assistant.

Frequently asked questions

How long should you wait before following up?
Long enough that something has actually changed since the first message — a date got closer, a fact arrived, or you narrowed the ask. Interval matters far less than content. If nothing has changed, waiting another week is better than sending a bump, because a bump spends the reader's attention and hands them nothing new to act on.
Is it rude to follow up more than once?
Not by itself. What reads as rude is a sequence with no visible end and nothing new in it. Two or three follow-ups that each carry new information and a narrower ask are easier to receive than one that only repeats the question, provided you say what you will do when the last one passes and then do it.
What should you do if someone never replies at all?
Take the silence as an answer, carry out whatever default you already stated, and tell them once, briefly, what you did. Non-reply is usually about the state of their queue rather than about you or your message, so the useful move is to remove the obligation rather than to escalate.