retention-architect: the flow map
What it does
Section titled “What it does”Finds the retention revenue you are not collecting, and specs the flows that would collect it.
It audits your existing email and SMS automation coverage and its share of revenue against the benchmark that flows should be carrying a large share of total email revenue, then prices every gap in revenue-per-recipient headroom rather than listing missing flows abstractly. A missing browse-abandon flow is not a checkbox; it is a number.
Then it produces 8 to 12 complete flow build specs: trigger, splits, timing, per-message copy direction, discount logic, compliance fields. Ranked by revenue gap, so you build the expensive one first.
Two disciplines it holds to:
- It reads the send-readiness verdict from deliverability-doctor before speccing any volume. Speccing sends into a restricted list is how you turn a retention project into a deliverability incident.
- Customer language only at the rung the evidence supports. A banked verbatim gets used as one; an unbanked but real line gets anchored to its source; anything else is labelled an operator hypothesis.
Read-only spec generator. The flows get built in your ESP.
When to reach for it
Section titled “When to reach for it”- “Which flows are we missing?”
- Retention revenue is flat and you want the ranked list of what would move it.
- Onboarding a client with an ESP nobody has audited.
- You need an abandoned-cart or post-purchase spec you can hand to whoever builds it.
Where it runs
Section titled “Where it runs”Both.
What it needs from you
Section titled “What it needs from you”- ESP access or an export: which flows exist, their revenue, and trigger-event volumes.
- The deliverability verdict, which it will ask for if it does not have one.
- The VOC bank, for copy direction grounded in real language.
Where a number genuinely is not available, that row ships marked NO-DATA with the missing input named, and anything derived from it is withheld rather than estimated. An unranked row is honest; a confidently ranked row built on a guessed number is not.
What you get back
Section titled “What you get back”- The flow map: what exists, what it earns, what is missing.
- Every gap priced in revenue headroom, with the arithmetic.
- 8 to 12 complete build specs, each with trigger, splits, timing, message-by-message copy direction, discount logic and compliance fields.
- A ranked build order.
A worked example
Section titled “A worked example”You: “Audit my flows.”
Coverage: 5 flows, 22% of email revenue against a benchmark near 41%. Biggest priced gap: browse abandonment, absent entirely; at your traffic and current revenue-per-recipient band the headroom models to roughly $4,100/mo. Second: post-purchase cross-sell, exists but single-message. Note: the win-back spec is capped at a lower send volume because the deliverability verdict is
SEND-RESTRICTED, and that constraint is carried into the spec verbatim.
Tips & gotchas
Section titled “Tips & gotchas”- Build in ranked order. The list is sorted by revenue gap, and the fourth item is usually worth a fraction of the first.
- Do not override a
SEND-RESTRICTEDconstraint. It came from a real deliverability read, and a retention push into a struggling list makes the list worse. - A
NO-DATArow is not a failed audit. It marks a number you can go get, and the spec ships without pretending. - Flows compound. Unlike ads, this revenue keeps arriving after you stop working on it.
Related skills
Section titled “Related skills”Requires deliverability-doctor’s verdict. Uses voc-miner for copy direction. Copy execution goes to hermes.