◆ Idea Factory
← All resources

Testing Business Ideas

David J. Bland, Alexander Osterwalder · 2019 · John Wiley & Sons

ExperimentsValidationStrategyzer

Overview

Testing Business Ideas is the practical companion that answers the question every entrepreneur faces after sketching a Business Model Canvas or Value Proposition Canvas: “How do I find out if any of this is actually true?” Written by David J. Bland — a Silicon Valley practitioner who has helped companies worldwide find product-market fit using lean startup and design thinking methods — and Alexander Osterwalder, co-inventor of the Business Model Canvas, the book provides a systematic, visual process for replacing assumption with evidence.

Published in 2019 in the same large-format, full-colour visual style as Business Model Generation, the book carries the Strategyzer series aesthetic and philosophy: practical tools over theory, visual over prose, team-ready artifacts over individual analysis. Its central premise is blunt: seven out of ten new products fail to deliver on expectations, and the primary reason is that organisations invest in building and scaling ideas before adequately testing whether those ideas have merit. The book sets out to reverse that ratio by making experimentation structured, repeatable, and accessible.

The audience is deliberately broad — solopreneurs validating a side project, startup teams managing investor expectations, and corporate innovation teams closing the gap between strategy and execution. All of them face the same challenge: they have ideas (often articulated on a Canvas) but lack a disciplined process for turning those ideas into evidence.

The core framework: Assumption Mapping and the Experiment Library

The book’s methodology rests on two pillars: knowing what to test (Assumption Mapping) and knowing how to test it (the Experiment Library).

Assumption Mapping

Before running any experiment, a team must make its assumptions explicit. Assumption Mapping is a collaborative exercise in which every uncertain belief embedded in a business idea is surfaced, written down, and then evaluated on two dimensions:

  • Importance — how critical is this assumption to the viability of the idea? If it turns out to be false, does the whole model collapse, or is it a minor inconvenience?
  • Evidence — how much supporting evidence already exists? Is this a guess, an informed belief, or a validated fact?

Plotting assumptions on this two-by-two grid immediately reveals the critical unknowns: assumptions that are both highly important and poorly evidenced. These are the ones that must be tested first. Testing assumptions with weak evidence but low importance is a form of procrastination — it generates activity without meaningfully de-risking the venture.

The Experiment Library

Once priority assumptions are identified, teams select experiments from a library of over 40 distinct experiment types. Each experiment is described with its purpose, how to run it, what evidence it produces, typical cost, typical time required, and the relative strength of evidence it generates.

The experiments are organised around the three dimensions that any viable business idea must satisfy:

  • Desirability — do customers want this? (discovery interviews, landing page tests, fake-door tests, surveys, customer advisory boards)
  • Feasibility — can we actually build or deliver it? (prototypes, concierge tests, technical spikes)
  • Viability — will the economics work? (pre-sales, letter of intent, revenue model tests, unit economics validation)

Experiments are further arrayed by their position on a spectrum from discovery (open-ended, generative — useful when you know little) to validation (specific, confirmatory — useful when you have a strong hypothesis to test with a clear success criterion).

The Test Card and Learning Card

Two lightweight artifacts structure each experiment cycle:

The Test Card makes a single experiment explicit before it runs. It captures: the hypothesis being tested; the experiment method chosen; the metric that will be measured; and the minimum threshold — called the criterion — that constitutes a pass. Specifying the success criterion in advance prevents the common failure mode of retroactively declaring success when results are ambiguous.

The Learning Card captures what actually happened: what was observed, what insight was drawn from that observation, and what action will be taken as a result. Together, Test Card and Learning Card create a tight learning loop — plan, execute, capture, decide — that mirrors the scientific method in a form any business team can use without prior research training.

Key concepts

  • Evidence-based decisions, not intuition-based bets. The book does not argue that experimentation replaces judgment; it argues that judgment should be grounded in evidence, and that evidence must be systematically gathered rather than anecdotally recalled.
  • Cheap before expensive. Experiments should be sequenced from lowest-cost/weakest-evidence to highest-cost/strongest-evidence. An interview costs an afternoon; a full product build costs months and capital. Run the interview first.
  • Pivot triggers. The Learning Card formalises the decision that follows each experiment: persevere (evidence supports the hypothesis), pivot (evidence suggests a different direction), or kill (evidence refutes the assumption fatally).
  • Organisational environment. The final section of the book acknowledges that individual experiments fail not only because of wrong assumptions but because of organisational cultures that punish failure. Leaders must actively create psychological safety and measure teams on learning velocity, not just output.

How to apply it to your blueprint

After filling in any framework — a Business Model Canvas, Value Proposition Canvas, or Lean Canvas — treat every box as a collection of assumptions. Rank them using an informal Assumption Map: which assumptions, if wrong, would invalidate the whole idea? Which have the least evidence?

Pick the two or three riskiest, least-evidenced assumptions. For each, write a Test Card: state the hypothesis precisely, choose the cheapest experiment that could test it, define a metric, and set the minimum criterion for success. Run the experiment. Fill in a Learning Card. Only then invest more time in the next layer of assumptions.

Return to the relevant framework after each learning cycle and update what you now know. The Canvas stops being a static snapshot and becomes a living map of validated beliefs and remaining open questions.

Strengths and limitations

The book’s greatest strength is its comprehensiveness as a reference tool. The experiment library gives teams language and structure for conversations that would otherwise be vague (“we should test this”) and makes selection deliberate (“we should run a concierge test because we need strong desirability evidence and can’t afford a full build yet”). The Test Card / Learning Card pair is immediately usable with no additional training.

Its limitation is that it operates one level below strategy: it tells you how to test assumptions but not how to prioritise between fundamentally different strategic directions. Teams can also fall into experiment theatre — running many small tests that never accumulate into a coherent view of whether the overall model works. The book acknowledges this and recommends a Progress Board to track cumulative evidence, but that tool is described briefly. Teams serious about experimentation at scale should pair this book with dedicated sprint-planning and portfolio management practices.

Key takeaways

  • Surfacing and ranking assumptions before experimenting saves time and prevents the busywork of testing low-stakes beliefs.
  • The experiment library provides over 40 structured options organised by cost, time, and strength of evidence — always match the experiment to the strength of evidence you need.
  • The Test Card disciplines teams to specify the success criterion before running an experiment, preventing post-hoc rationalisation of ambiguous results.
  • Desirability, feasibility, and viability must all be tested; most teams test desirability and ignore the other two until it is too late.
  • Experimentation is an organisational capability, not just a toolkit — leaders must create the conditions for honest learning.

How it maps to the Business Idea Factory

This book is the validation engine that makes the app’s frameworks actionable. Every canvas or framework in the app — Business Model Canvas, Value Proposition Canvas, Lean Canvas — produces a set of filled-in boxes. Testing Business Ideas provides the next step: converting those boxes into a ranked stack of assumptions and then running the right experiments to validate or refute them.

The app’s SWOT and Value Proposition Canvas outputs are natural inputs to an Assumption Map: SWOT weaknesses and threats often correspond to assumptions about competitive dynamics or operational feasibility; the Value Proposition Canvas’s customer profile (jobs, pains, gains) contains assumptions about customer behaviour that must be tested through discovery interviews or usability experiments. The app’s AI follow-up questions function as a lightweight equivalent of the Test Card’s hypothesis-specification step — prompting founders to be precise about what they believe and why, before they act on it.

References