The lifecycle stack: how we'd build it from scratch in 2026
If you were starting a new SaaS today, what would your email + lifecycle stack look like? Here's our opinionated take.
A surprising number of our customers ask us the same question on their first onboarding call: 'If you were starting fresh today, how would you set this up?' Here's the answer we wish we'd had when we were running lifecycle programs ourselves.
Stop thinking in tools, start thinking in surfaces
Most teams pick the email tool first, then bolt on a CDP, then a help desk, then a survey tool, then a referral product. By the time they've shipped their welcome series, they have seven SaaS subscriptions and no coherent picture of the customer.
Flip the framing. There are six surfaces in lifecycle: broadcast email, lifecycle automations, audience CDP, analytics, replies inbox, and AI assistance. Pick a tool that does all six in one place. Bolt on specialists only where the data flow is small or where the integration is bidirectional and rock-solid.
The five questions every lifecycle stack has to answer
1. Who is in my audience right now? Not yesterday, not at last refresh — right now. If your tool can't answer this in real time, you're flying blind.
2. What changes about them, and when? Events drive segments. Segments drive automations. Automations drive sends. If the chain is broken at any link, your lifecycle program is leaking.
3. What did we send them, and what happened? Not 'delivered' — what happened. Opens, clicks, conversions, revenue, replies. The full chain or nothing.
4. What didn't we send them, and should we have? Suppression rules, frequency caps, quiet hours, do-not-send lists — the inverse of sending is often the most consequential.
5. What would the next-best action be? This is where AI earns its keep. Not as a chatbox. As an operator that proposes what to do next and lets you approve.
A working example
Imagine a new SaaS team, 30k users, two-tier pricing, B2B. Here's what we'd build, in order of priority.
Week 1. Pick an event schema and start sending events to Mailapp. Don't over-engineer it — `page.viewed`, `feature.used`, `subscription.upgraded`, `subscription.cancelled`, `invoice.failed`. Five events get you 80% of the segmentation you'll need for a year.
Week 2. Set up your audience properties: plan, last_seen, weekly_active_count, total_revenue. These four properties are the spine of every segment.
Week 3. Build three segments: free-active, paid-engaged, paid-at-risk. Three lifecycle automations: free-to-paid nurture, paid onboarding, paid re-engagement. Don't build more than three for at least a month. Resist the temptation.
Week 4. Connect your warehouse for analytics reconciliation. This is the step every team skips. Don't.
Month 2. Add a referral or product-update broadcast. Use the same audience and segments. Don't pick a new tool.
Month 3. Now you can start being clever. Send-time optimisation. A/B tests. Branching automations.
The mistakes you'll make
You'll over-segment. We've all done it. 17 segments for 17 personas, none of them with enough volume to send to. Resist. Three good segments beats thirty bad ones.
You'll under-test. The biggest lift in any program is the subject line. A/B every send, even when it feels excessive. Stat-sig matters; eyeballing doesn't.
You'll drift on attribution. Pick one model, document it, defend it. The point is not which model is right — the point is to be consistent.
Closing
Lifecycle is a craft. The teams that win at it are the teams that treat it as one — with rigour, with documentation, with a willingness to be wrong and try again.
We built Mailapp to be the tool a serious lifecycle team would build for themselves if they had the time. If that's you, we'd love to meet.