01
Who this is for
The sponsor who funded the fix and deserves a straight answer
- — treated group
- — comparison group
Client-insight readout · DS-7+1 · assembled by the engine
The Receipt
We acted. Did it work? — answered with attribution honesty.
Illustrative — the chart is a designed canonical scene (the platonic form a confirmed story presents as), not a specific client’s data. With real data, the engine selects the matching scene and toggles the client’s reality against it.
02
The decision on the table
Did the intervention move the number — and can we honestly say it was us?
03
What we believed walking in
The discomfort of half-believing your own success story.
04
What the data shows
The stated target held against the after; the rival explanations surfaced, not buried; the claim sized to what survives.
05
What it means for you
An organization that can't tell whether its fixes work will keep paying for the ones that don't.
06
The move
Scale / adjust / stop — decided on evidence, with the attribution caveats attached.
- — Replay the bet as stated (no quiet goalpost moves)
- — Test the delta past the cutoff
- — Convict or exonerate the rivals — then decide: scale, adjust, or stop
07
Act / don't act
If ignored: Skip the receipt and every future budget argument is theology: the fixes that work starve while the ones with better stories feast. — If acted: A decision — scale, adjust, or stop — with caveats attached and credibility banked.
08
The honest close
What next? What else?
- — comparison-group design for the next intervention
- — another period of post data
- — pre-register the next bet's target