Compounding Curve Review¶
Diagnostic review — instantiates Compounding Advantage Flywheel Design
Reads the shape of the marginal-return curve across successive increments to tell a still-improving flywheel from one that has quietly flattened or begun to reverse.
A Compounding Curve Review is a periodic reading of one thing: the return to the next increment, plotted as the cumulative state grows. Its defining move is to look at the shape of the curve, not the height of the total — because a flywheel is only real while each added unit is worth more, cheaper, or easier than the last, and total growth can keep climbing long after that has stopped being true. The review is agnostic about why the curve bends; it does not certify a driver or a story. It answers a narrower, harsher question: is this loop still compounding, has it plateaued, or has it turned over?
Example¶
A B2B SaaS company believes it has a compounding expansion flywheel: each cohort of customers lands small and grows, so revenue from existing accounts snowballs. Headline numbers look great — total revenue is up 40% year over year. The Compounding Curve Review ignores the total and plots the return to the next increment: net revenue retention by cohort, and the incremental margin on each newly onboarded account.
The curve tells a different story. Net retention, once 128%, has drifted to 103% over six quarters — expansion is barely outrunning churn. And the incremental margin on new accounts is falling, because the newest cohorts need heavier support to stay. The total is still rising only because gross new-logo volume is masking a loop that has flattened. The review's output is a single annotated chart: "per-increment return has decayed from strongly increasing to roughly flat, and support cost per new account is now rising — the compounding claim no longer holds at the margin." That reframes the roadmap debate from "how do we grow faster" to "what would restore the return on the next customer, if anything."
How it works¶
- Choose the increment and its return. Decide what "one more unit" is (a customer, a cohort, a transaction, a contributor) and what return matters (value, cost, quality, speed), then measure the return to that next unit rather than the running total.
- Plot against cumulative state. Order the increments by how much stock existed when each was added, so the horizontal axis is accumulation, not calendar time — a growing curve on a calendar can hide a flattening curve per unit.
- Read the second derivative. Classify the shape: still accelerating, linear, diminishing, saturating, or reversing. The interesting signal is the change in slope, which is where a healthy loop and a dying one first diverge.
- Flag the turn. Watch for the inflection where increasing returns become diminishing — the saturation point — and for the sign flip where an added unit starts subtracting value.
Tuning parameters¶
- Increment definition — what counts as "the next unit." Coarse increments smooth out noise but hide early turns; fine increments catch turns sooner but are noisier.
- Return metric — which return you track (marginal value, marginal cost, marginal quality). Track the wrong one and a curve can look healthy on price while rotting on quality.
- Cohorting — whether increments are grouped by vintage. Cohorting separates a maturing loop from a diluting one, at the cost of thinner samples per cohort.
- Smoothing window — how much you average before reading the slope. More smoothing steadies the trend but delays detection of a genuine reversal.
- Review cadence — how often you re-plot. Frequent reviews catch turns early; infrequent ones save effort but risk discovering saturation a year late.
When it helps, and when it misleads¶
Its strength is that it separates real compounding from vanity growth: a rising total is easy to celebrate, but only the marginal-return curve reveals whether the loop is still doing the work.[n1] It is the earliest honest warning that a flywheel has saturated, because per-unit return turns before the total does.
Its failure mode is reading the curve too literally over too short a window. Increasing-return loops are lumpy — a genuine flywheel can show a flat or dipping stretch (a new-cohort dilution, an infrastructure investment not yet paying off) that a jumpy reviewer mistakes for saturation, prompting a premature retreat from a loop that was about to re-accelerate. The classic misuse is choosing the flattering increment: plotting the metric that still rises while quietly dropping the one that has turned. The discipline that guards against this is to fix the increment and return metric before looking, track more than one return (value and quality, not just revenue), and require a sustained multi-period turn before calling saturation.
How it implements the components¶
marginal_return_curve— the review is this curve made explicit: the measured return to each successive increment plotted against accumulated stock, the archetype's falsifiable test of whether returns are actually increasing.saturation_and_reversal_monitor— by watching the slope's inflection and sign it flags the moment increasing returns become diminishing or negative, the trigger to shift from acceleration to governance.
It does not certify why the curve bends — the learning-by-doing driver and its lesson-capture (increasing_return_driver, learning_capture_repository) belong to its nearest twin, Experience Curve Review. This review reads the outcome curve; that one certifies the cause.
Related¶
- Instantiates: Compounding Advantage Flywheel Design — it supplies the marginal-return reading that tells the design whether the loop still compounds.
- Consumes: Data Flywheel Dashboard or another instrument supplies the raw per-increment metrics this review interprets.
- Sibling mechanisms: Experience Curve Review · Bubble and Lock-In Red Team · Cumulative Reputation System · Data Flywheel Dashboard · Fixed-Cost Amortization Plan · Open Standard or Portability Rule · Platform Seeding Program · Reinvestment Cadence
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: The review examines measured marginal returns against cumulative state and classifies the curve as accelerating, linear, diminishing, saturating, or reversing, producing a bounded diagnosis of loop health.
Nearest alternative: Analysis, Modeling & Optimization — Calculating slope and curvature supplies the evidence, but the mechanism's output is the reviewed finding about whether compounding remains healthy or has turned.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Economics & Finance
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Marginal economic analysis supplied the discipline of reading returns to the next increment rather than cumulative totals.
Related originating lineages:
- Innovation & Entrepreneurship — Lean-startup and growth practice contributes flywheel diagnostics, cohort dilution checks, and suspicion of vanity metrics.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Independent reviewer agreement; medium confidence.
Notes¶
[n1] Vanity metrics — Eric Ries's term (in The Lean Startup) for numbers that rise reassuringly but do not reflect the mechanism that creates value, such as cumulative totals that only ever go up. The marginal-return curve is the deliberate antidote: it isolates the return to the next unit precisely because the running total is the easiest number to be fooled by. ↩