Innovation Accounting
A measurement discipline for judging progress in uncertain ventures through assumptions, cohorts, learning milestones, and investment decisions.
Download Markdown (.md)Soroush’s Product Wiki · https://wiki.soroush.uk/topic/innovation-accounting/
01What it means
Innovation accounting evaluates whether an uncertain product or venture is becoming more viable when conventional revenue forecasts and delivery measures are misleading. It establishes a behavioral baseline, tests changes against explicit hypotheses, and uses actionable measures to decide whether to continue, alter, scale, or stop. Its purpose is accountable learning, not inventing friendlier metrics for innovation teams.[1][2]
- 01State assumptionsCustomer value and sustainable growth
- 02Set measuresBehavior linked to a decision
- 03Establish baselineSmallest valid product or experiment
- 04Test improvementCompare trajectory, guardrails, runway
- 05DecidePersevere, pivot, scale, pause, stop
The measures must be actionable and cohort-based where appropriate—not friendlier vanity metrics for uncertain ventures.
02Why it matters
Early initiatives can produce activity without progress: features ship, cumulative registrations rise, and presentations improve while core behavior remains unchanged. Innovation accounting connects evidence to a model of value creation. It gives teams and funders a common basis for staged investment and reduces dependence on charisma, vanity metrics, or premature business cases.
03When to use it
Use it for new products, business models, or transformation initiatives where uncertainty is high and lagging financial results cannot yet distinguish progress from motion. Mature products still need experimentation, but stable unit economics and operating metrics may provide a more direct accountability system.
04How it works
- State the leap-of-faith assumptions behind customer value and sustainable growth.
- Choose cohort-based behavioral measures and define what decision each can inform.
- Establish an honest baseline with the smallest valid product or experiment.
- Run changes intended to improve the model, separating effects where practical.
- Compare trajectories with pre-agreed thresholds, guardrails, and runway.
- Hold a structured decision review: persevere, pivot, scale, pause, or stop—and record why.
05A practical example
A workplace learning venture avoids celebrating total enrollments. It tracks activation by acquisition cohort, completion of an applied task, manager-observed use after thirty days, renewal intent, and delivery cost. After two iterations improve enrollment but not applied use, leaders revise the customer and workflow assumptions rather than buying more traffic.
06Where it breaks
No metric set can remove ambiguity from sparse, delayed, or confounded data. Teams may game bespoke measures or demand statistical certainty that small ventures cannot provide. Quantitative evidence should be combined with qualitative inquiry, and investment thresholds should reflect ethical risk as well as growth potential.
Common mistakes
- Renaming ordinary dashboards without linking metrics to hypotheses.
- Using cumulative totals that hide cohort behavior.
- Changing success thresholds after observing results.
- Measuring learning claims without changing an investment decision.
07Sources
- Why the Lean Start-Up Changes EverythingHarvard Business Review · Management review articleWhy the Lean Start-Up Changes Everything
- Testing Business IdeasStrategyzer · Creator book overviewTesting Business Ideas