Onboarding email sequence for signups who never activated

An onboarding email sequence is the mail you send after someone creates an account and before they reach first value. The job is activation: get them far enough into the product that it actually works for them. Whether they open the email is not the job.

Two things it is not:

• Not a welcome email. A welcome celebrates the signup and lists every feature. It can be one short note. Onboarding is the sequence that keeps going only while activation is still missing.
• Not a day-1, day-3, day-7 drip. People finish setup at different speeds. A calendar that treats everyone the same will nag people who already activated and abandon people who stalled on one step.

If almost nobody is signing up yet, this is not where your time should go. An onboarding sequence does not invent signups. It gets more of the ones you already have to first value.

Most teams smash several jobs into one "welcome series." That is usually why the mail feels wrong.

Keep never-activated people out of win-back and out of coupon "we miss you" mail. Keep already-activated quiet users out of "finish setup" mail. Share one frequency cap across every program so they cannot stack in the same inbox.

More subjects and bodies live in onboarding email examples.

Four emails is enough for most self-serve products. Do not open with a welcome. Do not list features. If they activate, cancel the rest. Simple products can run fewer.

These examples use a made-up company called Acme. Each one is from Alex at Acme to a user named Maya Chen. None of them offer a discount.

For more annotated bodies, see onboarding email examples.

Every rule in this guide rests on one decision you have to make first: which event counts as activation. It is the exit condition for the sequence, the outcome you score it against, and the reason someone leaves the audience. If you cannot name it, none of the rest of this runs.

The temptation is to pick the end of your setup checklist. That is usually the wrong answer. The action that feels important to your team is often not the action that predicts whether someone is still here in three months. You find the real one by looking at the people who stayed.

Work through it in this order:

1. Decide what retained means for your product. Thirty days after signup is a reasonable starting point. Products with a long value cycle may need ninety.
2. Split your past signups into the ones who were still active at that mark and the ones who were not.
3. List what a new account can do in its first week or two: invited a teammate, created a first project, connected a calendar, pasted a snippet, imported real data.
4. For each one, compare how many of the retained group did it against how many of the group that left did it. The action with the widest gap between those two numbers is your candidate.
5. Check that enough people currently miss it. An action almost every signup already takes is a fine milestone and a useless target, because there is nothing left to move.

Some products need a count and a window rather than a single action, such as a few teammates invited in the first ten days. Use a threshold only when the gap between the two groups actually widens at one, not because a well-known company published a number.

Then check that the event causes retention instead of just marking it. Someone may be inviting teammates because they were already going to stay. Use the same holdout you would use on any campaign: nudge a random half of never-activated signups toward that action, send the other half nothing, and see whether the nudged group retains better. If it does not, you found a symptom rather than a lever.

Two things worth keeping separate:

• The step you name in the email is not always the event you measure. The email names the specific thing they stopped on, because that is what makes it worth reading. Activation is the outcome you judge the sequence on. Those are often different, and that is fine.
• The aha moment is not the event. The aha moment is when someone realizes the product is worth it. The activation event is the behavior that shows it happened. If people are clicking through your steps without getting there, you are measuring compliance instead of value.

Re-check the gap every few months. Once activated and never-activated accounts retain at about the same rate, the event has stopped predicting anything and you need to derive it again. For the product side of this, see Activation for PLG SaaS.

A calendar drip assumes everyone who signs up moves at the same pace. They do not. Some people finish setup in ten minutes. Others take three weeks.

Fire the sequence from the signup event. Send the next note only while they still have not activated. Cancel everything the moment they do.

You need two events in the product at minimum:

1. Signup (or account create)
2. Activation (the first action that means the product worked for them)

Name the incomplete step from what you already know: opened invite modal, created an empty project, never connected a calendar, never pasted a snippet. Do not say "getting started."

Calendar timing is only a fallback while activation is still missing. It is not the reason someone is in the audience.

Quiet users who already activated belong on the lifecycle email guide. Cancelled accounts belong on win-back email.

The sequence has to stop on its own. That means an exit, skips for steps they already finished, and a frequency cap it shares with everything else you send.

• Cancel the rest on activation. The moment they do the thing, the remaining sends stop.
• Skip steps they already finished. If email 2 is about inviting teammates and they invited someone yesterday, that send should never go out.
• Do not stack automations. If another automation already covers this moment, fix the overlap instead of running both.
• Share one frequency cap. Onboarding, campaigns, retention, and win-back all land in the same inbox, so count every program against the same limit per person.
• Start with a small audience. Expand only the versions that beat a holdout.
• Honor a reply of "stop." Suppress marketing onboarding for that person.
• Hold the send when you emailed them in the last week, when you have nothing specific to say, or when complaint rates are climbing.

An always-on automation is fine after a version has beaten its holdout and you have approved it. Automation is for repeating something you already decided is good, not for inventing the sequence in the first place.

Build the audience from what you already know about each signup:

• When they signed up, and whether they have activated since
• The last setup step they started but did not finish
• The last thing they did in the product, if anything
• Which plan they are on, or whether they are still in a trial

These are too weak to trigger on by themselves:

• Everyone on day 3
• Anyone who has not opened an email
• The whole list

None of those tell you whether this person still needs this email.

Feed this from your analytics tool, such as PostHog, Amplitude, or Mixpanel, or from your own product events. The audience needs to update on its own as people activate. If you export a list once and mail it a week later, you will ask people to finish steps they have already finished.

A good test: if you cannot name the event that put someone into the sequence, that segment is not ready to send to.

Score the sequence on the activation event you picked earlier, or on paid conversion when that is the outcome you actually care about.

Then hold back a random slice of the same never-activated audience and send them nothing at all. Compare the two groups on activation or paid conversion. The gap between them is what the sequence caused.

Open rates cannot answer this question. Someone can open every email and never touch the product again.

If a version does no better than the holdout, stop sending it. If it wins, you can turn it into an always-on automation for people who sign up later, and get that approved before it goes live.

• You do not have meaningful signup volume yet. Getting more people to sign up is the bigger problem.
• You are building an ecommerce welcome or coupon series. Those run on a different set of events.
• The people you want to email already activated and went quiet. That is retention on lifecycle email.
• The people already left after real use. That is win-back email, not onboarding.
• You need one-time passcodes, account verification, or password resets. Those are transactional emails.
• You want one sequence that every signup gets forever, no matter what they do. That is the thing this guide argues against.