# SignalCraft Research Data Standard

Canonical: https://signalcraftinsights.tech/research-data-standard

SignalCraft will apply this standard to future aggregate product and research statistics.

Every published aggregate claim should state:

1. the population and unit counted;
2. the cohort and time window;
3. the exact outcome definition;
4. sample size, inclusion rules and exclusions;
5. any classification method and model version;
6. uncertainty, missing data and important bias;
7. privacy and aggregation treatment;
8. a review date and correction history.

## Platform action is not customer demand

Sessions, clicks, idea revisions and generated landing pages describe behaviour inside a founder platform. They do not by themselves establish that a customer has a problem, will purchase, will adopt or will remain.

## Current boundary

SignalCraft does not currently publish an outcome-calibrated startup success benchmark. It will not relabel product engagement as market validation or imply that an assessment score predicts commercial success.

Read [How to Read a Startup Validation Dataset](https://signalcraftinsights.tech/blog/how-to-read-a-startup-validation-dataset) and the [Evidence and Claims Register](https://signalcraftinsights.tech/evidence).