Client Zero

We ran the method on ourselves first. This website is the first thing we studied. Its claims are tiered, you can trace where each one comes from, and its machine representation is public.

Publishing our own research changes the thing we're studying: our objects can end up in what evaluators read and repeat back. We don't pretend that isn't happening. We measure it. Propagation Records track which of our objects show up in evaluator behavior, how accurately, and with what effect.

A company that is also a participant should say so. We do.