1. Frame the question
The idea, target customer, problem, geography, current alternatives and decision to be made define the research brief.
Research methodology
How SignalCraft turns public evidence into a market assessment
SignalCraft uses AI to search and organise evidence, not to invent market data or predict success. Every assessment is designed to make the sources, assumptions, contradictions and remaining uncertainty visible.
The research process
2. Plan the search
Queries are generated across problem pain, workarounds, competitor dissatisfaction, willingness to pay and direct requests.
3. Collect public evidence
SignalCraft searches current public discussions, reviews, communities, competitor material and other relevant sources.
4. Filter aggressively
Off-topic, duplicated and low-quality results are removed. Accepted sources are classified by relevance and evidence type.
5. Evaluate signal strength
Recurring behaviour, explicit requests, spending, workarounds and switching carry more weight than generic mentions or supportive opinions.
6. Assess commercial risk
Demand is considered alongside regulation, platform dependency, reputation, adoption resistance and switching costs.
7. Separate evidence from interpretation
Reports show the supporting sources, explain the reasoning and state where evidence is missing or contradictory.
8. Recommend the next test
The output identifies what the current evidence justifies and which uncertainty should be tested next.
Evidence hierarchy
Actions that require time, money, switching or repeated effort are stronger signals than likes, survey enthusiasm or a single complaint.
- • Repeated behaviour and paid alternatives
- • Workarounds and switching attempts
- • Explicit requests and active searches
- • Opinions and hypothetical interest
Limitations
- • Public evidence is not primary customer research.
- • Niche B2B markets can have sparse public discussion.
- • Evidence reflects a point in time.
- • No report can determine acquisition cost or retention before testing.
- • A strong problem does not automatically create a viable business.
Inspect the output
The sample report shows the evidence categories, confidence drivers, commercial risks, unknowns and next experiment produced by this methodology.