2023 — Customer Retention Rate was assumed stable under Decay Over Time in Healthtech during Post-Funding

Why did our healthtech retention rate look strong in the first six months — and then deteriorate steadily without any product change?

A digital health platform raised a Series B in early 2023 on the basis of 12-month retention figures that appeared competitive with category leaders. The retention model used in investor materials measured cohorts only to month six. When 18-month and 24-month cohort data was later compiled, the retention curve showed a structural decay pattern that began after the initial behaviour-change motivation wore off. The platform had been measuring engagement during the honeymoon period and projecting it forward as if it were a steady state. By the time the decay was visible, the company had hired to a headcount plan and built a product roadmap that assumed the flawed retention base.

Failure Type:
→ Assumption Failure

Crux:
→ Permanence Illusion

Variable Hub:
→ Customer Retention Rate

Case

The platform offered personalised chronic condition management programmes. Early adopters were highly motivated — they enrolled following a health event or doctor referral, which created an initial engagement spike. Month 1–6 retention sat at 74%. The company presented this cohort slice in its Series B deck. What the deck did not show: month 7–12 retention was 51%, and month 13–24 retention was 31%. The decay was a known property of health behaviour change programmes, where motivation is event-driven and habitual behaviour is difficult to sustain without structural intervention. No decay model had been built into financial projections.

Decision Error

Post-funding, the company built LTV projections using the 6-month retention rate extrapolated forward on a flat curve. Payback periods, pricing strategy, and enterprise contract lengths were all calibrated to this projection. The assumption was never stress-tested against a decay scenario, despite the team’s awareness that health behaviour programmes typically exhibit declining engagement curves.

Why It Failed

Health behaviour change follows a well-documented motivation arc: acute event triggers enrolment, initial compliance is high, and attrition accelerates once the triggering urgency fades and habit formation fails. Retention in months 1–6 measures crisis response, not long-term product value. Using it as a proxy for steady-state retention is a category error. The decay is not a product failure — it is a structural property of the user population that must be modelled explicitly and addressed through re-engagement design.

Trigger

The trigger was a quarterly business review in month 14 post-raise, at which the finance team presented LTV actuals against model. Actual LTV was 44% below projection. The source of the gap was traced to retention decay that had been outside the model’s scope. By this point, the company had committed to three enterprise contracts priced on the flawed LTV assumption and had hired a 40-person clinical team to support projected user volume that did not materialise.

Missed Signal

Academic literature on digital health engagement programmes consistently documents retention decay after month 6. Comparable category companies had disclosed churn data in due diligence materials that showed the same pattern. Neither the founding team nor the lead investor had modelled the decay curve before signing the term sheet. The signal existed in public domain research — it was never integrated into the planning framework.

Rule

If stability is assumed, test for change before committing.

Decision Criteria (Machine Logic)

IF ALL conditions below are TRUE:

Retention rate is measured on a short cohort window and treated as representative of long-term behaviour
Decay mechanism from motivation fade is not modelled in LTV or financial projections
Headcount and contract commitments increase dependency on the projected retention figure
Enterprise contracts and pricing are locked before 18-month cohort data is available
commitment exceeds rollback threshold

THEN → Permanence Illusion

Failure Pattern

Ontology Pattern:
Temporary Condition → False Stability → Commitment → Exposure → Failure

Variable Pattern:
High early retention (motivation-driven) → Flat retention curve assumed → Funding and contracts committed → Decay surfaces post-month-6 → LTV actuals miss projections → Enterprise repricing required

Outcome:
Two enterprise contracts were renegotiated at a 30% price reduction. One was terminated. The clinical team was reduced from 40 to 22. The company reoriented its roadmap toward re-engagement features that should have been built before the initial contracts were signed.

Intervention

Build retention decay curves from category benchmarks before any LTV projection is used in planning
Stress-test LTV at month-12 and month-24 retention rates, not month-6
rollback if threshold exceeds 30 days

If validation is not possible → Do not proceed.

→ The Decision Ledger
→ Assumption Failure
→ Permanence Illusion