The Nuance Collective · 2026

The Nuance Index™

How clearly do AI assistants understand a category? Every Friday we measure one — whether the answers they give are settled, accurate, and anchored to the real market. One locked method, all year.

How to read it →

Rationale

What it measures

The subject is the category

The Index asks whether AI assistants give a coherent account of a whole category — settled, accurate, and anchored to the real market. One category a week, the same instrument every time.

Clarity, measured three ways

Whether it answers the same way twice, whether what it says is true, and whether its ranking matches the real category. Those three roll into one number from 0 to 100, and a word for which kind of clarity it is.

Brand positions are the evidence

Where each brand lands is how we read the category. How clear the category is then decides whether that position can be moved, which is why the verdict matters more than the row.

The method in plain English →

How to read it

Six things worth knowing

It scores a category

The number describes the category's clarity. Where a brand lands is evidence of that, which is why the same position means different things in a settled category and an unsettled one.

Not a visibility metric

Share-of-voice tools count how often a brand appears. Clarity asks whether the picture it is appearing inside makes sense at all. A brand can be highly visible in a category the model gets wrong, and no visibility number can tell that apart from a good position.

Known ingredients

We publish the dimensions, the component scores, the reference set, the run count and the exact category nouns. We never publish the weighting or the prompt wording.

Bands, not places

Twenty-five runs are a sample. Every figure carries the range it moves inside when those runs are resampled, and brands are grouped into bands cut where the ranges stop overlapping. Inside a band, the data does not separate them.

A stated inclusion rule

Each edition publishes what qualified an entity to be measured — entity type, size band, classifying authority, as-of date, source linked. The thresholds are ours and labeled as ours; the entity list and the figures are not.

Comparable only to itself

A 62 in snack crackers does not mean what a 62 in another category means. Each edition stands alone, and no edition cites another's score.

The full guide — two minutes →