# Validated Learning

Decision-relevant learning supported by credible evidence that tests a consequential assumption about customers, value, or viability.

- Type: principle
- Domain: Strategy & discovery
- Level: Advanced
- Last reviewed: 2026-09-22
- Original: [Soroush’s Product Wiki](<https://wiki.soroush.uk/topic/validated-learning/>)

## What it means

Validated learning is a demonstrated reduction in uncertainty about whether a product direction can create sustainable value. It requires more than collecting feedback or reporting activity: the team states a prior belief, obtains evidence through an appropriate method, considers rival explanations, and changes confidence or action. Validation is always bounded to the population, context, measure, and time observed.

## Why it matters

Teams can learn interesting facts without reducing strategic risk. Validated learning focuses inquiry on beliefs that determine investment and makes progress possible before financial outcomes mature. It also creates an auditable connection between research, experiments, and decisions, helping organizations avoid both opinion-driven persistence and reflexive reactions to noisy data.

## When to use it

Use the principle when developing products, entering markets, or changing business models under substantial uncertainty. It is especially useful for governing staged funding. For established obligations—security patches, accessibility compliance, or contractual work—learning may improve execution but is not the primary justification.

## How it works

1. Specify the consequential assumption and current confidence before gathering evidence.
2. Define what observations would increase, decrease, or leave confidence unchanged.
3. Select methods suited to the claim: interviews for meaning, behavioral tests for action, and longitudinal data for persistence.
4. Check sampling, instrumentation, ethics, and plausible alternative explanations.
5. Update the assumption and record what remains unknown or context-bound.
6. Tie the learning to a concrete decision, next test, or change in investment.

## A practical example

A team assumes independent restaurants abandon inventory tools because setup takes too long. Observation confirms setup friction, but a concierge test that completes setup for owners does not improve four-week use. Interviews reveal that unpredictable supplier substitutions are the deeper issue. The original problem claim is weakened, and the next investment shifts accordingly.

## Where it breaks

Evidence rarely validates an entire strategy, and repeated testing can still reinforce biased assumptions if samples and methods remain narrow. The language of validation may imply final truth or encourage binary verdicts. Teams should report degrees of confidence, transfer limits, negative evidence, and unresolved uncertainty.

## Common mistakes

- Calling any customer conversation validated learning.
- Testing a proxy that is weakly connected to the strategic assumption.
- Ignoring evidence that does not fit the preferred narrative.
- Claiming universal validation from a narrow context.

## Sources

1. [Why the Lean Start-Up Changes Everything](<https://hbr.org/2013/05/why-the-lean-start-up-changes-everything>) — Harvard Business Review · Management review article
2. [Testing Business Ideas](<https://www.strategyzer.com/library/testing-business-ideas-book>) — Strategyzer · Creator book overview

## Related topics

- [Build–Measure–Learn](<https://wiki.soroush.uk/topic/build-measure-learn/>)
- [Innovation Accounting](<https://wiki.soroush.uk/topic/innovation-accounting/>)
- [Hypothesis-Driven Design](<https://wiki.soroush.uk/topic/hypothesis-driven-design/>)
