Why UMP exists

§1InformativeDraft 2.3Latest draft

This section is informative.¶

UMP lets a test result, a model and a budget plan pass between tools without losing what the numbers mean, so two analysts asking the same question reach the same answer or can see exactly why they differ.¶

The case that started it is a Thursday before budget lock. A geo test on Instagram Reels reads out 30% below what the MMM predicted. The planning lead wants next quarter’s plan to reflect it. The MMM came from a vendor as a spend-response table, so nobody can refit it. One analyst copies the Meta curve into Excel and scales it down by the test result. A second analyst, working on a sister brand, scales the whole Paid Social curve instead. The client receives two recommendations $400K apart, and a four-day audit cannot reconstruct either.¶

Frame 1. The readout

The MMM predictedA vendor spend-response table. Nobody can refit it.

The Reels geo test read30% below that prediction

It is the Thursday before budget lock. The planning lead wants next quarter’s plan to reflect the test.
Frame 2. Two fixes

Analyst one Copies the Meta curve into Excel and scales it down by the test result. Recommendation A

Analyst two, sister brand Scales the whole Paid Social curve instead. Recommendation B

$400K apart The client gets two recommendations. A four-day audit cannot reconstruct either. Bar lengths are illustrative.
Frame 3. Why the checks passed

Suppose both analysts had used the first draft of UMP perfectly: shared names, model tiers, logged data changes. The numbers still answer different questions.
- Compared with what?The vendor table measured ROI against zero spend. The test measured a 30% cut.
- Over how long?The Reels lift covered two weeks. The curve assumed full carryover.
- Schema checksEvery one would pass.
The fix is in the contract. A number now crosses UMP with its estimand, an uncertainty that combines honestly, its provenance and its trust level. Mismatched estimands are never combined without a stated assumption (G-EST-02).

Every number was right on its own. Combined, they were wrong, and nothing in the files said so. That is the gap UMP closes.

Figure 1.1 The Thursday before budget lock: two analysts, $400K apart. Step through the story in three frames: the test readout, the two fixes, and why every schema check would still pass. The bars are illustrative; only the $400K gap comes from the story.

The first draft of UMP fixed the plumbing: shared names, model tiers, a fallback ladder, logged data changes. The research review then asked a harder question. Suppose both analysts had used UMP perfectly. Would their numbers have agreed? Not necessarily. The vendor table measured ROI against zero spend, the test measured a 30% cut, and the Reels lift covered two weeks while the curve assumed full carryover. Every schema check would pass and the numbers would still be wrong to combine. The same review found the acceptance bar would have passed a model with the same flaw as the Media Tracking Model, because it tested whether the code ran correctly and never whether the model could see causes at all.¶

So this draft treats meaning as part of the contract. A number crosses UMP with its estimand, its uncertainty in a form that can be combined honestly, its provenance and its trust level. When two pieces of evidence do not line up, the gate engine gives the best answer it can still defend and says what it had to assume. A decision made on a labelled best guess beats one made in an unlabelled spreadsheet because the system said no.¶

What success looks like, measured against Phase 0 baselines:¶

Measure	Target
Test readout to an approved plan¶	Three working days or less
Client-facing numbers that resolve to an artifact ID¶	90% by end of Phase 3; 100% for work at the Full materiality tier
Unexplained conflicting answers to one standard question¶	Zero per quarter
Mismatched estimands combined without a stated assumption¶	Zero (`G-EST-02`)
Stale scenarios quoted after a model changes¶	Zero
Newton exact-answer accuracy on the golden set¶	Proposed 90% or higher, with 100% citation validity