How to Write a Welcome Email Sequence That Engages
What the first five emails have to accomplish, in what order, and the timing that matches how quickly people forget signing up.
On this page
The short version
- The first email is read at a rate nothing else you send will match. Spending it on a greeting rather than on delivery wastes the only guaranteed attention you get.
- Send the first one immediately. The gap between signing up and hearing from you is the window in which people forget who you are.
- Five emails is a ceiling, not a target. A three-email sequence that is read beats a seven-email one that is abandoned at four.
A welcome sequence is the only email you write that reaches people at the exact moment they decided you were worth hearing from. Everything after it competes with a reader who has forgotten the decision. That makes it the highest-leverage writing in the whole programme, and it is routinely the oldest, least-revisited thing in the account.
The sequence has five jobs, and they map cleanly onto five sends. If you have fewer than five, they merge — they do not disappear.
The five jobs, in order
Timings assume a content or software list. A retail list can compress this; a long-cycle B2B list can stretch it.
Deliver what was promised — immediately
Whatever the form said they would get, this email gives it, above the fold, as the first thing. No preamble, no company history, no request.
Immediately means immediately. A delay of even an hour means a proportion of people have moved on, and some will have forgotten subscribing at all — which is where spam complaints come from.
Set the expectation — day one or two
What arrives, how often, and from whom. This is the email that prevents future unsubscribes, because most unsubscribes are a mismatch between what someone expected and what turned up.
It is also where you can invite a reply. A question that is genuinely worth answering produces replies, and replies are the strongest positive signal available to a mailbox provider.
Prove it — day three or four
Your single best piece of existing material, chosen because it demonstrates the promise rather than because it performed well on social. This is the email that converts a subscriber into a reader.
Position — day six or seven
Who you are and why you are the one writing this. This lands correctly here and badly in email one, because by now you have given three things and earned the right to talk about yourself.
Ask — day eight or ten
The first commercial ask, framed against everything already delivered. If your cycle is long, this becomes a soft ask — a reply, a call, a free tool — rather than a purchase.
Where a sequence loses people
The shape below is typical rather than universal, and the useful part is the ratio between steps rather than the absolute figures. Your own numbers will differ; the pattern of where the drop concentrates usually will not.
Illustrative shape for planning. Measure your own sequence and replace these — the point is that the largest single drop is between one and two, which is why email two carries the expectation-setting.
The gap between one and two
The biggest fall in almost every welcome sequence is from the first email to the second, and the reason is structural rather than editorial: email one is opened because the reader is waiting for something. Email two arrives unrequested.
That is why expectation-setting belongs at position two rather than position one. It is the email doing the most work to make position three get opened, and putting it later means three, four and five all inherit a weaker audience.
Branch on where they came from
Someone who signed up from a technical guide and someone who signed up from a pricing page want different sequences. Both should get the same five jobs; the proof email and the ask should differ.
This is worth building even at modest list sizes because it is a one-time cost. Two variants of two emails inside a five-email flow is an afternoon, and the flow then runs for years.
What each send is measured on
| Send | Primary measure | Failing when |
|---|---|---|
| 1 — delivery | Open rate | Below the high eighties on a confirmed list |
| 2 — expectation | Open rate versus send one | Drop worse than about forty percent |
| 3 — proof | Click rate | Opens hold but nobody clicks |
| 4 — positioning | Unsubscribe rate | A spike here means it reads as a pitch |
| 5 — the ask | Conversion, and unsubscribes | Conversion with a large unsubscribe spike |
Measure per send, not per sequence. A sequence-level conversion figure tells you it is not working and nothing about where.
Sequence review
- Email one sends within seconds, not on a schedule
- Nothing in the sequence names a price, feature or person that has changed
- There is an exit condition, and buying triggers it
- The expectation set in email two matches the actual cadence
- Someone has read all five in order, as a subscriber, in the last six months
- The reply address is monitored — email two asks for replies
Frequently asked questions
How long should the whole sequence run?
Eight to ten days for most content and software lists, compressed for retail, stretched for long-cycle B2B. The constraint is that it should finish before your regular cadence would have reached them twice, or the two streams start colliding.
Should the regular newsletter pause during the welcome sequence?
Suppress it, yes. Receiving a welcome email and a campaign on the same morning reads as two senders. Most platforms can exclude anyone currently in a flow.
What if people signed up before the sequence existed?
Do not retro-send it — a welcome email two years late reads as a mistake. Write a separate re-introduction if the list is genuinely cold, and treat that as a win-back rather than a welcome.