How to measure brand consistency
Measuring brand consistency does not require a research budget. It requires reading a brand the way a first-time visitor does and scoring whether the same story holds across surfaces. Here is a practical method, and how the Net Entropy Score automates it.
The problem: most brand reviews are not measurements
A typical brand consistency review is a workshop and a slide deck: useful opinions, but subjective, one-off, and impossible to compare or re-run. Two reviewers reach two conclusions, and neither produces a number you can track next quarter.
To manage consistency you need a method that is repeatable (same inputs, same score), explainable (you can see why), and comparable (you can run it on a competitor). That is a measurement, not a review.
Why checklists and audits fall short
Style-guide checklists measure the visual layer (logo, color, type) and miss the expensive part: whether the promise, audience, voice, and proof are the same across pages. A bigger brand audit goes deeper but is bespoke and rarely re-run, so it captures a moment, not a trend.
The gap is a standardised model: a fixed set of dimensions, scored the same way every time, from signals anyone can see.
A simple method you can follow
Read the homepage, the main product or service page, the pricing page, and the about page as a stranger. For each, note the single promise it makes, the audience it speaks to, the voice, and the proof it offers. Then look across them: does the promise stay the same, does the audience stay the same, does the proof back the claims, and does any surface contradict another.
Where the answers diverge, you have found inconsistency. Counting and weighting those divergences is, in essence, what a brand consistency score does, just systematically and from more signal.
This manual method approximates the NES website layer.
Automating it with NES
NES (Net Entropy Score) runs that method automatically. It reads a brand's public pages, scores ten consistency dimensions, and returns a 0 to 100 brand consistency score with the reasoning and evidence for each, in about two minutes. Because the model is fixed, the score is repeatable and comparable.
For a fuller read, the Review-Inferred tier measures consistency from public customer language, and the Measured tier from a direct cohort survey.
The ten things it scores
The model scores five components that build coherence and five that subtract from it. NES is the net of the two, expressed 0 to 100 for the website layer. Each is a directional reading on one dimension of consistency, drawn from the brand's own surfaces.
As a benchmark, across 119 DTC websites in June 2026 the typical score was 71 and the highest 80, so the low 70s is roughly the current bar to beat.
Frequently asked
How do you measure brand consistency?
Read the key pages as a first-time visitor and check whether the promise, audience, voice, and proof stay the same across them, then weight the contradictions. The NES framework does this systematically and returns a 0 to 100 score with evidence.
Can I measure brand consistency for free?
Yes. A free website scan returns a 0 to 100 brand consistency score with a component breakdown in about two minutes, no login required.
What should I read to assess consistency?
At minimum the homepage, the main product or service page, the pricing page, and the about page. Deeper reads also bring in public reviews and news to compare the claim against the reality.
Is a single score enough?
The score is the headline; the component breakdown is what you act on. It shows which specific dimensions and pages are pulling the number down.