How the diagnostic is scored

Each engine scores seven structural dimensions, forty-two across the six-engine platform. Each dimension is measured by four anchored questions, and the four are not interchangeable. Two are objective and two are perception.

Objective anchors ask what is measurably true of the business: what proportion, how many days, how often, what is documented. Perception anchors ask how the operator experiences the same territory. Both are answered on a five-point scale where every point is defined in words, not left to interpretation.

The two kinds are then weighted differently. Objective anchors carry sixty percent of a dimension's intake and perception anchors carry forty. That split is deliberate: it lets the instrument detect the gap between what is true and what is believed, which is frequently where a structural constraint hides, while still letting measurable fact dominate the score.

This is what "calibrated" means here. It is a fixed instrument with defined scale points, the same for every business that takes it, so two companies with the same structure receive the same score regardless of who reads the intake.

Scores resolve into four bands. The thresholds are fixed in the scoring kernel, not chosen per engagement:

A band is a structural reading, not a grade. A business can be Structurally Robust on most dimensions and still be badly constrained, because the constraint is set by the binding dimension, not the average.

Each dimension carries a fixed weight in the composite score, and the weights sum to exactly one. They are set per engine and do not move between engagements.

This matters for reading your own report. A weak score on a heavily weighted dimension moves the composite far more than the same score on a light one, so the composite is not an average of equals. The report shows the per-dimension breakdown alongside the composite for exactly this reason: the composite tells you the altitude, the breakdown tells you where the weight is sitting.

Compound patterns are also detected across dimension pairs. Two dimensions that are individually survivable can form a structural pattern that neither score shows on its own, and those pairs are evaluated explicitly rather than left for a reader to infer.

The report is not assembled freehand. Every volume is built against a fixed set of section contracts, each with its own completeness requirements, so no section can silently ship thin:

Any figure that reaches a client report has to clear one of three origins, and nothing else is permitted:

One. Arithmetic on the numbers you supplied. Your own inputs, computed transparently.

Two. A public structural fact about your sector, cited to its source so you can open it yourself.

Three. An explicitly designed band, labelled as designed and pending calibration, never presented as an observed measurement.

The phrase "industry research shows" is not permitted anywhere in the pipeline. If a number cannot name which of the three it is, it does not ship.

This is also why peer benchmarking is described rather than displayed today. Comparing your structure against a peer cohort is only defensible once enough comparable diagnostics exist to make the comparison mean something. Until then it is held rather than shown. A benchmark presented before it is defensible is a decoration, and it is the specific thing this discipline exists to prevent.

Stated plainly here rather than left in the engagement terms for you to find.

The scoring is deterministic. Your anchor responses run through the same kernel every time, and the scores, bands, constraint logic and pattern detection are computed, not authored. A language model does not decide your score.

The narrative prose is generated by a language model working from those computed results, under a set of automated quality gates that check the writing against the scoring: whether every claim traces to an anchor, whether sections repeat each other, whether the recommended moves are specific enough to act on. Volumes that fail those gates do not go out.

A human reviews the volume before delivery. The instrument is the reason two clients with the same structure get the same reading. The review is the reason a machine does not have the last word.

A consulting engagement that locates a structural constraint typically runs into five figures and takes weeks, and what you are buying is a specific partner's judgement. The judgement is often excellent. It is also unrepeatable: a different partner, a different week, a different answer.

What is being sold here is the opposite shape. A fixed instrument, a fixed scale, fixed weights, a fixed report contract, and a provenance rule that governs every number. It costs less than the consulting engagement because the analysis does not have to be rebuilt from nothing each time. It costs more than a questionnaire because there is a scored kernel, a gated report pipeline and a human review behind it, and because the output is a document you can act from rather than a result page.

The honest limit: this locates structure. It does not implement the fix, and it does not promise what the fix will be worth to you.

It does not promise a business outcome. No revenue change, no cost reduction, no operational improvement is committed to, predicted, or implied. The diagnostic tells you where the structure binds. What happens next is your execution, your market and your timing.

It does not benchmark you against a peer cohort yet, for the reason given above.

It does not audit your finances or verify your inputs. The instrument reads the structure you describe. If the intake is answered carelessly, the reading reflects that.

It is not advice. It is a diagnosis. It tells you what is true of the structure, not what you should decide.

Any example or range shown anywhere on this site is illustrative and is not a commitment or a prediction.

How Crown Mosaic finds your binding constraint walks the process from intake to artifact. This page covers how the scoring underneath it works.

What Crown Mosaic Platform is covers the category and the six engines.

Pricing lists every tier and what each includes.

Glossary defines the terms used above.

Engagement terms carries the full honest-proof statement, the AI disclosure, and the satisfaction guarantee that applies to the single-engine OPERATE diagnostic.