Turn data into a higher standard
Power BI.
Uncompromised.
Review. Redesign. Verify.
Turn ordinary Power BI reports into decision-grade products.
- Cleaner modelsOne definition per number
- Stronger decisionsBuilt around the decision
- Evidenced findingsNothing on taste alone
- Stated limitsWhat a review cannot prove
Beneath the surface
Every number stands on four layers.
- 01The surfaceIt looks finished. Most reports do.
- 02UnderneathVisuals stand on measures, measures on a model, the model on its sources. Take the report apart and you can see what holds it up.
- 03The findingsEach layer is read against The X Standard. What does not hold is marked where it lives, with its evidence.
- 04VerifiedRebuilt around the decision, then checked that the meaning survived the work.
Illustrative data · the Margin Review
The report, rebuilt
Now it’s yours. Move it.
The Margin Review from above, rebuilt around its decision as widgets. Drag any card to rearrange the page; on a touch screen, press and hold first; with a keyboard, focus a card, hold Alt and use the arrow keys.
Drag to rearrange. On touch screens, press and hold first. With a keyboard, hold Alt and press the arrow keys.
Illustrative data · the same story the hero and the film tell
What PlayWithData X is
A report becomes a product when someone can act on it.
Most Power BI work stops at “the numbers are on the page”. PlayWithData X carries it the rest of the way: review what is already there, redesign it around the decision it is meant to serve, build it properly, and verify that the meaning survived the work.
- 01
It starts with the decision, not the dataset
The first question is which decision the report exists to support, and who personally owns it. A report nobody has acted on in ninety days has a finding, not an audience.
- 02
Definitions are treated as the product
Two measures can carry one name and mean different things. Agreeing what each number means is the work; the visual is the packaging.
- 03
Findings carry evidence, not taste
Every finding points at something checkable — a definition, a relationship, a key coverage, an unanswered question — so it can be argued with.
- 04
The limits are stated out loud
Some things a review can measure. Some it cannot. Saying which is which is what makes the rest worth trusting.
The X journey
Review. Redesign. Verify.
Four stages, in order. Each one is listed with what it actually is today, not what it will be. Nothing below is available to buy.
xReview
Prototype · not availableA structured assessment of the report you already have: the decision it is meant to support, the definitions behind its numbers, the model underneath it, and whether anyone has acted on it. Produces evidenced findings and, only where the evidence carries one, a banded xScore printed with its evidence tier and coverage. Never a single number, and nothing at all when the evidence cannot support one.
Prototype: 8 screens, an annotated specimen, and a scoring engine whose integer output is superseded by the band contract · Specified: rubric, intake contracts, golden and adversarial tests · No backend, mocked data
xDesign
Prototype · not availableRedesign around the decision: a leading question instead of six equal tiles, a hierarchy that puts the argument first, and a path through the report that ends somewhere a person can act.
Prototype only: a redesign screen inside the review prototype, against one hand-built specimen · No method specified
xBuild
Prototype · not availableThe implementation itself — model, measures and report delivered as a working artefact with a blueprint behind it, not a picture of one.
Prototype only: an implementation-blueprint screen in the review prototype · Separately, four real PBIR/TMDL mutation scripts exist as benchmark evidence, untested and fixture-bound
xVerify
PlannedEvidence that the meaning survived the work: model, measures, visuals and the decision they support still agree — and a plain statement of what a check of this kind cannot prove.
Adversarial tests and golden reviews drafted · No verification engine exists in any repository · Estimated 30–40 engineering days
Status is read from the repository, not from ambition. A stage moves to “available” only when something runs.
Honest status
One prototype. One standard. No customer.
A runnable review prototype exists: eight screens, a working scoring engine, an annotated specimen and a redesign, built to be judged rather than sold. It has no backend, its data is mocked, and it has never been run against a real client report.
Alongside it sits The X Standard v0.1: a rubric, a scoring approach, intake contracts, golden reviews and adversarial tests. The two were written a day apart and do not yet agree with each other. Nothing here is for sale, and no customer has been through any of it.
Status as at 2026-09-11 · Sources: prototype delivery report · X Standard v0.1 consolidated report · Report data shown in the film is illustrative
The method
What can be claimed, and what cannot.
Two pages carry the method: the standard a report is read against, and the premium expression of the journey. Neither is settled, and both say so.
Request a review
Start with the question.
Begin where a review begins: name the report, and name the decision it is supposed to support. The rest follows from those two answers. A person reads every request.