blueNull manual

Audits realized A/B counts for achieved minimum detectable effect and integrity issues (sample-ratio mismatch, exposure dilution), and signs the result — so a “not significant” dashboard reading can be checked against whether the test could have detected anything at all.

What this product is and is not

Is: Could the experiment have seen the effect?.

Is not: Not a causal-inference platform. A NULL_LICENSED verdict does not claim the underlying effect is exactly zero.

Who uses it

Experimentation and analytics reviewers.

Getting started — the local demo

From the product's own repository:

make demo
make demo-ui   # to see the same scenarios in the web dashboard

This runs entirely offline against synthetic fixtures and exits non-zero on any mismatch. It does not connect to any production data source.

Screen-by-screen walkthrough

Audit history — powered vs. underpowered results, and sample-ratio-mismatch flags

What a receipt proves

A signed receipt proves what blueNull ran and recorded; it does not prove an underlying real-world claim is true. Anyone can check the signature offline with the standalone verifier shipped alongside the product, without installing blueNull itself.

Known gaps, stated plainly

Not a causal-inference platform; no sequential testing or CUPED in V1.

FAQ

Does blueNull make the decision for me?

No. Not a causal-inference platform. A NULL_LICENSED verdict does not claim the underlying effect is exactly zero.

Does blueNull replace the system I already use?

No — it is an overlay. Your existing record system, LIMS, or claims platform stays exactly where it is.