The Lab
Base data. Sized opportunity. Proven lift. We don’t guess at what’s losing you customers — we find it, put a dollar figure on it, and fix it continuously.
Tier one
ONE-TIME · 1–2 WEEKS
The agent team reads your analytics, flags every point of friction with the mechanism behind it, puts a dollar figure on each, and ranks the list. You get the backlog and the reasoning behind it — yours to keep, whether or not we work together again.
Tier two
RETAINER · MONTHLY
The full loop, running: diagnosis, test design, simulation, execution. Lift reported monthly. Every test builds your Materia, so the tenth test is designed by a team that remembers the first nine.
The loop
We ingest your analytics and session data and flag where people hesitate, backtrack and leave — each one tied to a named behavioral mechanism rather than a chart that dips.
Every finding is priced. Not “this could be significant” — an expected annual revenue figure, with the assumptions visible so you can argue with them.
The highest-value findings become real tests with a hypothesis, a target metric, a sample size and a stopping rule fixed before launch.
Results ship, get killed, or come back null — and all three write to your Materia with their confidence attached, so the next test starts from what you know.
Anatomy of a finding
38% of sessions that reach the delivery step change nothing and then abandon. The preselected option is the most expensive of three. A default with no stated rationale reads as a sales tactic, and hesitation at this step is consistent with that reading.
Illustrative. Every figure on a real finding links back to the query that produced it.