The five-email sequence everyone has
Day 0: welcome. Day 1: here are our features. Day 3: a case study. Day 5: book a demo. Day 7: your trial is ending.
Every B2B SaaS company has some version of this, it's written once, and it goes out on a timer regardless of whether the user has done anything at all.
Which means a user who set up the product, imported their data, and invited two colleagues receives the same "getting started" email as someone who signed up and never logged in again. One of them is being told to do something they did four days ago. The other is being sent a case study when they can't get past the first screen.
Both emails are ignored, and the sequence gets judged on open rates that tell you nothing about either failure.
The reframe
An onboarding email is not a marketing message that happens to arrive during onboarding. It's a part of the product that runs outside the interface — the mechanism that brings a user back when the product itself can't reach them.
Once you hold that framing, three things change.
It should react to state, not to time. The product knows what the user has and hasn't done. A sequence that ignores that information is choosing to be less intelligent than the system it's part of.
Its job is to move someone to the next step, not to inform them. The measure isn't opens or clicks. It's whether the user came back and did the thing.
It's owned jointly with whoever owns activation. Written in isolation by marketing, it will describe features. Written with product, it addresses where users are actually getting stuck — which is knowable from the data.
What a state-based sequence looks like
Instead of five emails on a timer, you have a small number of branches keyed to where the user stopped.
Signed up, never returned. The most common and most valuable segment. They haven't seen anything yet, so a feature tour is useless. One email: the single fastest path to one specific outcome, with a realistic time estimate. "Import a file and get your first report in about 10 minutes."
Started setup, didn't finish. They hit friction at a known step. Address that step specifically — the most useful thing here is usually removing work rather than explaining it. A template, sample data, or an offer to do the import for them.
Set up, hasn't reached first value. Configured but hasn't experienced the thing. This is where a well-placed example or a short guided path earns its keep.
Activated. Now — and only now — the messages that assume comprehension make sense: advanced features, other use cases, inviting the team, case studies from similar companies.
Four branches, each triggered by a product event. Considerably less work than most teams expect, because most of the content already exists and is simply being sent to the wrong people.
The practical requirements
An activation event. You cannot branch on state without defining what state means. This is the same prerequisite as almost everything else worth doing in a funnel, and it remains the most common thing missing.
Product events reaching your email tool. The usual blocker, and it's infrastructure, not strategy. If your product analytics and messaging platform can't see each other, none of this is implementable regardless of how good the plan is.
Suppression that actually works. If someone completes the action, they must stop receiving the prompt. Getting an email telling you to do something you did yesterday is worse than no email — it tells the user the company isn't paying attention.
How to measure it
Not opens. Not clicks.
Did the user perform the target action within 48 hours of receiving this message?
Compare against a holdout that received nothing. Most teams discover that two or three of their emails do all the work and the rest are noise that costs them unsubscribes.
That's an uncomfortable finding and a useful one. A three-email sequence that reacts to what the user did will beat a nine-email sequence that reacts to the calendar, and it takes less time to maintain.
Hilal Tasdan
B2B SaaS Growth Marketing Consultant & Fractional CMO. Partner in Growth.