swan-gtm Applies to any subscription product with usage data. Produces an audit ranked by leverage, or a staged roadmap with data prerequisites named.
Audit the data layer before anything else
The most common finding is not a missing campaign. It's that the events and attributes the messaging would need are not being tracked, are tracked inconsistently, or never reach the messaging platform. Design a beautiful behavioural program on top of that and it silently sends the wrong thing to everyone.
So the first move in any engagement is: what events actually exist and are populated? Are plan, usage, and subscription attributes flowing through? Where's the gap between what the messaging needs and what the data provides? If the pipes are broken, fixing them is the first deliverable — say so plainly, even when the ask was for a campaign.
The eight stages
Locate the user on this map, then ask: what's the single next value moment, and what message helps them reach it?
- Acquisition — signup created → set expectations, drive first login
- Activation — first login → reach the aha moment
- Onboarding — account created → complete the setup steps that predict retention
- Adoption — activated → build the habit, widen feature use
- Retention — recurring use → reinforce value, prevent decay
- Expansion — hitting limits → upgrade, seats, cross-sell
- Advocacy — promoter → referral, review
- Resurrection — dormant or churned → reactivate, save cancellers, recover failed payments
A minimum viable program set is: welcome/activation, onboarding, trial conversion if there's a trial, failed-payment recovery, and win-back. Add expansion and advocacy once those are working.
Two truths that shape every stage
Retention is the foundation. Acquisition, expansion, and referral all leak away without it, and top-line growth hides decaying cohorts for a surprisingly long time. Read the cohort retention curve before prescribing anything: a curve that flattens means there's a real product-market fit to build on; a curve decaying toward zero cannot be fixed by email, and saying so is the honest deliverable.
Behaviour beats time. Trigger on what someone did or didn't do at their natural frequency, not on a fixed clock.
Define activation properly
The activation metric is the earliest milestone that predicts long-term retention — the aha moment repeated at the expected frequency, not a setup step completed once. A validated activation metric shows activated users retaining at roughly twice the rate of non-activated ones. If yours doesn't, you've measured a setup step and every program keyed to it is aimed at the wrong target.
What good looks like
The tell of a good operator: they ask what the core action is and how often a happy user would do it, before asking anything about email. Everything downstream — triggers, delays, what counts as dormant — falls out of that one answer.
The mediocre version is a complete-looking set of flows built on a fixed time drip, with no holdout, no exit conditions, and a churn number nobody can attribute to anything. It looks like a working lifecycle program in a screenshot.
Good output ranks gaps by leverage against this company's actual churn shape rather than listing every missing program, states the data prerequisite for each recommendation, and instruments measurement plus a holdout before launch — because a program you can't attribute is a program you can't defend or improve.
Rules
- MUST audit event and attribute coverage before designing any program.
- MUST define the core action and its natural frequency first.
- MUST rank recommendations by leverage, not by completeness of the stage map.
- MUST instrument metrics and a holdout before launch.
- NEVER prescribe messaging as the fix for a retention curve decaying toward zero.