Turn market observations into a testable thesis with a mechanism, disconfirming evidence, alternatives and an expiry date before choosing channels or campaigns.
Documented fact: Formal evaluation guidance recommends defining a theory of change, questions and evidence before judging outcomes.
CDM inference: The Falsifiable Market Thesis Canvas adapts that logic for marketing decisions; it is CDM synthesis and has not been externally validated as a standard.
A market thesis is a decision claim, not a trend summary
A useful thesis explains how a change in a defined market is expected to affect a defined customer decision and therefore the organisation’s choices. “AI is changing marketing” is a topic. “Procurement teams will require verifiable provenance for AI-assisted product claims, increasing the value of auditable content operations in regulated categories” is a thesis that can be investigated.
The thesis should have consequences. If it is true, the team should allocate attention, evidence and capability differently. If it is false, the team should change course. A statement compatible with every possible outcome is positioning language, not a decision instrument.
The Magenta Book’s evaluation principles emphasise clear questions, theory and proportionate methods. CDM’s canvas brings that discipline into market planning while recognising that most companies cannot run a clean experiment on an entire market.
Complete the Falsifiable Market Thesis Canvas
CDM asset: Falsifiable Market Thesis Canvas
- Market boundary: geography, category, segment, decision and period.
- Observed change: what has been measured or directly reported.
- Mechanism: why the change should alter behaviour.
- Prediction: what should appear first if the mechanism operates.
- Business implication: the decision that becomes more or less attractive.
- Alternatives: at least two explanations that fit the observation.
- Disconfirming evidence: what would reduce confidence.
- Expiry: the date or event that forces re-evaluation.
Start narrow enough to be wrong. “Among UK software buyers with formal procurement, evidence requests will appear earlier in evaluation over the next two quarters” is easier to test than “trust matters more now”. The boundary prevents evidence from another segment or country being used opportunistically.
Write the mechanism as a chain. For example: more AI-assisted content increases claim uncertainty; procurement responds by requesting provenance; suppliers with accessible evidence clear evaluation faster. Each link may fail. That is an advantage because it shows where to collect evidence.
| Canvas element | Weak version | Decision-grade version |
|---|---|---|
| Observation | Customers are cautious | Security and evidence questions now appear before demonstrations in a defined opportunity cohort |
| Mechanism | Trust is important | Earlier procurement involvement raises the proof threshold before technical evaluation |
| Prediction | More content will help | Accounts receiving a complete evidence pack should reach approved evaluation more often, with no rise in unsuitable leads |
| Disconfirmation | We will know from results | The sequence does not differ after evidence access, or interviews identify budget freeze as the binding cause |
Build evidence that can disagree with you
Use three evidence layers. Market evidence describes category, regulation, competitor behaviour and economic context. Customer evidence captures decisions, alternatives and language through interviews, observation and behavioural records. Company evidence shows pipeline, product use, margin, fulfilment and retention.
Separate the source from the interpretation. A customer saying “we need to wait” is a recorded statement. Concluding that price caused the delay is an inference until the conversation or behaviour supports it. A competitor launching a lower-cost tier is a fact; claiming it will reset the category reference price is a forecast.
Create an evidence ledger with date, source, segment, method, finding, limitation and thesis implication. Give contradicting evidence equal visibility. A team that only records supporting evidence is producing advocacy, not market intelligence.
Interview non-buyers, former customers and people who chose an alternative. Existing customers can explain adoption, but they are a selected group. Include distribution and frontline evidence without treating sales opinion as a substitute for direct research.
Choose tactics as tests of the mechanism
Once the thesis is explicit, ask which intervention creates the clearest learning while contributing to the business. If the predicted problem is proof friction, build a source-linked evidence package and test whether suitable buyers use it and progress. A broad awareness campaign would be a poor first test because multiple mechanisms could explain the result.
Map every tactic to a predicted behaviour and a decision. A webinar is not a strategy. It might test whether procurement teams engage with a live technical explanation. A comparison page might test whether buyers need a transparent category frame. Define the signal, the guardrail and what the team will do under three results: support, ambiguity or contradiction.
A tactic can succeed while the thesis fails. Strong creative may generate response even when the proposed market change is absent. Likewise, a tactic can fail because execution was weak although the thesis remains plausible. Evaluate implementation, mechanism and outcome separately.
Run a thesis review before the story hardens
Set a review date when the thesis is written. Bring the evidence ledger, predicted leading signals and alternative explanations. Ask a reviewer who did not sponsor the thesis to present the strongest case against it.
Use four outcomes: retain when the mechanism has credible support; revise when the boundary or sequence changes; split when segments behave differently; or retire when disconfirming evidence is stronger. Record the decision and its consequences for budget, content and capability.
Do not rewrite the original thesis after the result. Version it. Preserving what the team believed and why makes learning possible and prevents hindsight from turning every outcome into confirmation.
Limits and risks
Markets are open systems. Competitor moves, regulation, supply and culture can change during the test. Evidence may support a mechanism locally without establishing it across the category. Treat the thesis as a bounded working explanation.
Falsifiability does not make weak data strong. A precise claim based on biased interviews remains fragile. Match research effort to decision risk, protect participant privacy and seek specialist help for regulated or technically complex markets.
CDM verdict
A market thesis earns strategic value by making the team vulnerable to being wrong. Define the boundary, mechanism, prediction, alternatives and disconfirming evidence before channels enter the room. Tactics should create value and learning; they should not be used to protect a fashionable story.
Related guides
Frequently asked questions
What is the difference between a market insight and a market thesis?This article is editorial guidance. Apply the principles in proportion to your market, evidence, and responsibilities.
