The Cohort Payback Path Review Framework

OBSERVED PAYBACK PATH — original Growthcraft review worksheet
Decision / owner / review date:
Acquisition cohort entry event / source / version / count:
Customer identity reconciliation and test-order exclusions:
Currency / tax convention / channel attribution rule:
Acquisition spend scope and allocation / finance approval:
Bucket duration / customer-age boundaries / timezone / snapshot:
Last fully observed bucket for the entire cohort:
Refund convention: posting age or restated original-order age:
Variable costs included / credits and recoveries / missing costs:

1. RECONCILE — analyst
Frozen customer denominator, including non-repeaters:
Revenue, refund and cost totals versus source exports:
Empty completed bucket distinguished from missing data:
Evidence IDs and unresolved differences:

2. TRACE — analyst and finance
Contribution by bucket = revenue before refunds - refunds - variable costs
Cumulative contribution / allocated acquisition cost:
First observed crossing / any subsequent reversal:
Final endpoint covered? / last below-cost endpoint:
No interpolation or forecast supplied by this worksheet.

3. REVIEW — growth owner
Comparison cohort at the same fully observed age:
Product, channel, pricing and acquisition-cost mix differences:
Would missing refunds or costs change the conclusion?
Separate cash timing / causal evidence / risk budget review:

4. DECIDE — approver
Hold for missing evidence / document observed result / propose bounded investigation:
Next action, owner, evidence required and review date:
No automatic channel scaling or universal payback target.

When this review is useful

This original operating framework is for D2C, consumer subscription and startup growth teams that have an acquisition cohort and successive contribution observations. It answers whether a recovery claim survives the data contract and later refunds. It is not a validated score, investment recommendation or claim that an attributed channel caused those customers.

Use it after the same starting customers have reached each included observation boundary. Do not use it to fill immature months with zeros, compare a new cohort's first week with an old cohort's full year, or convert revenue ROAS directly into recovered acquisition cost.

Inputs and owners

The analyst supplies cohort membership, extraction versions, bucketed revenue, refunds and variable costs. Finance defines acquisition spend, cost allocation and credits. The growth owner states the decision and comparable cohort; an approver owns any change in spend. A small startup can combine roles, but not silently skip the checks.

Use anonymised aggregates in the worksheet. Retain source-level identities only in your controlled data environment. Define whether refunds appear at posting age or restate the originating order; these conventions answer different questions. Name the latest fully observed age for the entire cohort and preserve zero-activity members.

A four-stage decision sequence

1. Reconcile: match membership and totals to source exports. Unresolved cost coverage or a changing denominator means hold, not an adjusted score.
2. Trace: compute incremental bucket contribution, its cumulative sum and every crossing of allocated acquisition cost. Preserve negative buckets and disclose reversals.
3. Review: compare like-aged cohorts, cost definitions and product mix. Separate observed recovery from predicted future orders, marginal acquisition efficiency and cash settlement.
4. Decide: record the narrow supported conclusion and a named next action. A reconciled path can justify further investigation, not automatically a bigger budget.

The arithmetic boundary is contribution equal to acquisition cost. It is a definition, not a commercially sufficient target. Risk appetite, overhead, financing and opportunity cost require a separate review.

Worked reversal example

Synthetic, EUR. Allocate 5,000 acquisition cost to 100 customers. Completed 30-day buckets contain revenue/refunds/variable costs of 4000/500/2000, 6000/500/2000, 200/700/100 and 2000/100/700. Contribution is 1,500, 3,500, −600 and 1,200. The cumulative path is 1,500, 5,000, 4,400 and 5,600.

The first observed crossing is bucket 2, but bucket 3 reverses it. At bucket 4 the cohort again covers acquisition cost, with a 600 balance. The worksheet should record both facts, not report “permanently paid back in 60 days.” If finance has not reconciled the refund file, keep the decision on hold even though the arithmetic is correct.

Failure modes and boundaries

Averaging only repeat buyers inflates recovery. Deducting acquisition spending inside contribution and again at the threshold double-counts it. Truncating negative buckets hides refund effects. Carrying forward missing months silently invents evidence. An endpoint crossing gives no exact event date, and a later order or chargeback can change the path.

Shopify describes a monthly gross-profit approximation; Glencoyne discusses channel-specific acquisition costs and contribution. This worksheet adds an explicit observed-path and reversal review, not a vendor-endorsed benchmark. Shopify payback guide; Glencoyne channel payback guide.