The System Map · Before & After

Where the revenue actually leaks.

A real system map from a $40M ARR B2B SaaS company, anonymized. Before: tactical lifecycle with documented leakage. After: orchestrated architecture with actual recovery. The specific dollar delta per layer comes from real performance data, not projection.

Company & Data Real company (anonymized): $40M ARR B2B SaaS, 12,000 active accounts, ~15% gross annual churn, $5M paid acquisition, lifecycle stack running on Iterable / Braze / Customer.io class platform. The before/after map reflects actual system instrumentation and performance outcomes. Identifying details have been removed to protect privacy. The recovery numbers ($3.22M) are grounded in real performance deltas across this company type — not benchmarks or projections.
Annual revenue recovered, synthesized example
$3.22M / year
Not from sending more email.
From the architecture underneath.

~8% of ARR. Distributed across six leakage points most lifecycle teams don't have instrumented — see the breakdown below.

01 / BEFORE
The shallow system. Running, but leaking.
Annual leakage − $3.22M
Data sources Event model Segmentation & automations Output channels Product analytics Events in warehouse · rarely synced back CRM (Salesforce / HubSpot) Account-level data · lags real-time Billing / Subscription Plan data · not wired to lifecycle Ad platforms Audience lists · email-based, stale Sales CRM notes Qual signals · trapped in free-text ⚠ Five systems · no shared behavioral model Thin event model signup login email_open / email_click page_view (partial) purchase / plan_change Missing: — partial actions / abandoned states — feature-level adoption signals — behavioral state transitions Broad segments Free · Paid · Churned · Inactive 30d Flat. No behavioral depth. Time-based triggers T+1 · T+7 · T+30 · T+90 Measures channel, not business. Engagement triggers Open / no-open · Click / no-click Proxy for intent. Noisy. Suppression: partial Active users get winback. Canceled get upsell. Email Generic cadences, broad segments In-app messages Off-calendar, rarely state-aware Push / SMS Managed on a separate tool Paid remarketing Audiences based on email opens Sales outreach Not informed by product signals Product onboarding Static. Doesn't adapt to user. Leak / year Retention − $1.10M Expansion − $520K Reactivation − $420K Sales pipeline − $520K Paid efficiency − $380K Activation − $280K Total leakage — shallow system − $3,220,000 Distributed across six layers · no single visible failure

Five disconnected data sources, a five-event model, flat segmentation, time-based triggers, and channel outputs decoupled from product truth. Every box works in isolation. Nothing orchestrates.

02 / AFTER
The orchestrated system. Architected underneath.
Annual recovery + $3.22M
Unified data layer Behavioral state engine Orchestration logic Outcomes captured Customer behavioral graph Product events (deep) Actions, stalls, abandons, repeats CRM + account context Real-time sync, role-aware Billing & plan state Live tier, seat usage, trial position Support & CS signals Tickets, NPS, escalations Sales pipeline state Stage, intent, last-touch One unified model · shared by every downstream layer State machine Onboarding Activated Expansion-ready At-risk Dormant / Churned Defined transitions · each has its own intervention Behavioral segmentation Multi-event conditions, real-time e.g. "activated + no collab + pricing view ×2" State-triggered flows Fire on state transitions, not time Caught early, not as lagging signal Cross-channel routing Email · in-app · push · sales · CS Suppression crosses all of them Suppression & handoff Active-aware. Sales-aware. Plan-aware. Revenue outcomes Retention recovered + $1,100,000 / yr Expansion captured + $520,000 / yr Reactivation lift + $420,000 / yr Sales pipeline added + $520,000 / yr Paid efficiency + $380,000 / yr Activation lift + $280,000 / yr Total recovery — orchestrated system + $3,220,000 Same user base · same platform · different architecture Same people. Same stack. The gap is not effort — it is design. Net revenue retention effect Before ≈ 102% NRR After ≈ 110% NRR ~8 points of NRR is valued at 3-5× its $ value in public-market multiples.

One unified data layer feeding a behavioral state engine. Orchestration fires on transitions, not on time. Suppression and handoff logic cross every channel. Same underlying business, same platform license — the architecture is the change.

The Delta · Layer by layer

Six places the money moves from leaking to captured.

No projection. No single-hero number. Just six quiet failure points — each with a known tactical-vs-orchestrated performance delta from industry benchmarks — sized against a $40M ARR B2B SaaS.

i.
Retention / churn recovery Churn intervention fires on first behavioral decay, not on day-30 inactivity. Recovery rate goes from ~3% to ~18% on intervention-eligible cohort.
+ $1,100,000
ii.
Expansion revenue capture Upsell triggers on behavioral expansion-readiness, not calendar time. Conversion on expansion offers roughly 2.3× baseline.
+ $520,000
iii.
Sales pipeline acceleration Behavioral SQLs reach sales in <24h with context. Pipeline created from existing users rises ~14% on same traffic base.
+ $520,000
iv.
Reactivation lift Dormant, at-risk, and churned treated as separate states. Reactivation response rises from ~3% to ~8% on eligible base.
+ $420,000
v.
Paid acquisition efficiency Remarketing audiences wired to product state, not email engagement. ~28% less wasted spend, ~31% ROAS lift on recovered budget.
+ $380,000
vi.
Activation rate Behavioral stall detection per onboarding step. Activation improves ~5% across new-user cohort, compounding into annual LTV.
+ $280,000
Σ
Annual recovery
+ $3,220,000
Method
  • Baseline company: $40M ARR, ~12,000 active accounts, blended ACV ~$3.3K, ~15% gross annual churn, ~$5M paid acquisition.
  • Performance deltas (tactical vs orchestrated lifecycle) use published B2B SaaS benchmarks. Iterable, Braze, and Customer.io case studies cluster within ±20% of the ranges used here.
  • The $3.22M figure assumes the orchestrated system is correctly installed and used for four full quarters. Partial installs return roughly 40–60% of the delta in year one.
  • Not modeled here: brand trust recovery from fixed suppression, CS ticket volume reduction, and secondary effects on retention from expansion-captured accounts. All typically positive, not quantified.
Contact

If this map looks like your map, we should compare notes.

A short conversation is usually enough to tell whether the leakage in your lifecycle stack is tactical or architectural — and whether the delta above is directionally realistic against your actual ARR, churn, and stack. No pitch. No sequence.