hub
Research
We measure things and publish the method alongside the result. Nothing here is published until its sample meets the minimum we set in advance.
What belongs here
This section holds original measurement only: cohorts, markets and metrics, each with a methodology version, a sample size and stated limitations.
A result is published only once its sample meets a minimum that was defined before the measurement ran. A run that falls short is recorded and withheld rather than published with a caveat, because a caveat does not stop a number being quoted.
Nothing is published yet
This section is deliberately empty. A benchmark is published only once its sample meets a minimum defined before the measurement ran, and we do not yet have a sample that meets one.
What will appear here is original measurement of website and software quality: cohorts, markets and metrics, each with its methodology version, sample size, measurement window and stated limitations.
Publishing a number from a sample too small to support it would be easier than saying this, and it would be a fabricated claim. The platform refuses to publish a benchmark whose sample failed its minimums, so the honest state of this page is that it has nothing to show yet.
How a result will be published
- A definition states the cohort, the metrics and the minimum sample before anything is measured.
- A run records exactly which targets it drew, so the figure can be re-derived from stored measurements.
- Every published figure carries its sample size, methodology version, window and limitations.
- A figure is never presented as a population-wide statistic; it describes the sample that was measured.