How We Build Brand Health Benchmarks From Real Audits

Every benchmark figure published on this site comes from audits people actually ran here. Nothing is modelled, estimated or borrowed from a third-party study. This page explains exactly how those numbers are produced, what we deliberately refuse to publish, and how much weight an industry average deserves when you compare your own score against it.

The short version: we take completed audits, group them by industry, suppress any group too small to stay anonymous, and publish averages plus the failures that recurred most often. No business is ever named, and no individual finding is ever republished.

What goes into a benchmark

A benchmark is built only from completed audits. Drafts, failed runs and abandoned intakes are excluded, because a partial audit produces partial evidence and would drag averages in a direction that reflects our infrastructure rather than the state of the brands we measured.

Each completed audit contributes:

• its overall weighted brand health score, out of 100
• its per-category scores across the scored categories
• which specific checks it failed

Nothing else. Business names, URLs, hostnames, email addresses and verbatim findings never enter the benchmark tables. The aggregate is the product; the source rows stay private.

Benchmark data component	Included in aggregate benchmark	Excluded / Suppressed
Overall brand health score	Yes (weighted score out of 100)	Incomplete or abandoned runs
Category scores	Yes (calculated across scored categories)	Individual draft evaluations
Check failures	Yes (if appearing in ≥25% of industry runs)	Rare checks (<25% frequency)
Identifying data (URLs, names, emails)	No	Always excluded from aggregation
Industry cohorts (<8 completed audits)	No	Suppressed entirely for anonymity

The minimum sample rule

An industry only gets published figures once at least eight completed audits sit behind it. Below that threshold, the section is suppressed entirely rather than shown with a caveat.

There are two reasons for this, and only one of them is statistical. The statistical reason is obvious: an average of three is a rumour. The other reason is anonymity. In a small enough sample, an unusual score becomes identifying — anyone who knows two of the three participants can infer the third. Suppression is how the promise that no business is named survives contact with small industries.

The same rule applies to the failure percentages. A specific failure is only reported when enough audits in that industry actually ran the check, and only when it appeared in at least a quarter of them. Rare failures are real, but they are not benchmarks.

What the figures mean, and what they do not

An industry average tells you where the middle sits. It does not tell you what good looks like, and it is not a target.

Read a benchmark as a diagnostic prior. If eighty per cent of audited brands in your sector fail the same check, that failure is not a personal embarrassment — it is a category-wide blind spot, and clearing it is unusually cheap differentiation. If your score sits well above the average but your enquiries have not moved, the benchmark has told you something useful too: the constraint is somewhere the audit measures well and you are already strong in, so look at demand rather than brand expression.

What this does not tell you

• Causation. A high brand health score is not evidence that brand health caused revenue. The audit measures brand expression and its discoverability, not sales outcomes.
• Your competitors specifically. The sample is whoever ran an audit, not a representative census of the industry. Early samples skew towards businesses motivated enough to audit their page experience and technical performance.
• Anything about a named business. Deliberately so.
• Trend, until enough time has passed. Period-over-period movement in a small sample is mostly sampling noise.

We would rather state the limitations plainly than let a number carry more authority than it earned. If you quote these figures, quote the sample size with them.

How the numbers stay current

Benchmarks are recomputed on a schedule as new completed audits arrive, so every published figure carries the sample size and the period it covers. When an industry's sample grows past the threshold for the first time, its section appears; if a period passes with too little data, it is suppressed again rather than carried forward stale.

You can read the current figures on the brand health benchmarks report, and see how they break down by sector on the industry brand audits hub.

The Brand Health Audit is a platform built by The Branded Agency. The benchmark work exists because the alternative — publishing individual customers' findings for search traffic — would trade their reputations for our rankings, and we were not prepared to do that.

Where to go next

The most useful comparison is your own. Run a free audit and read your category scores next to the published averages for your sector. If a category sits far below the benchmark, start there — the what to fix first guidance walks through how to prioritise.

Frequently asked questions

Where does the benchmark data come from?

From completed brand health audits run on this site. Each contributes an overall score, per-category scores and which checks it failed. No third-party datasets, industry reports or estimates are mixed in.

Could my business be identified in the benchmarks?

No. Business names, URLs, contact details and individual findings are never published, and any industry with fewer than eight completed audits is suppressed entirely so that small samples cannot be reverse-engineered.

Why do some industries show no figures?

Because they have not reached the eight-audit minimum yet, or the reporting period did not contain enough completed audits. We suppress rather than publish a thin number with a warning attached.

Is the average a target score?

No. It is the middle of the sample, not a standard of good practice. Use it to spot where you are unusually weak or unusually strong relative to your sector, then prioritise by commercial impact rather than by closing the gap to average.

How often are the benchmarks updated?

They are rebuilt on a regular schedule as new completed audits arrive. Every published figure shows the sample size and the period it covers, so you can see how fresh and how large it is.

Can I cite these figures?

Yes. They are free to quote with credit and a link back to the benchmark report. Please include the sample size alongside any figure you use.

Sources

• Creating helpful, reliable, people-first content — Google Search Central. Guidance on creating reliable, people-first digital content and establishing brand trust signals.
• Page experience in Google Search — Google Search Central. Details how technical performance, accessibility, and user experience factors are evaluated.
• The business case for digital accessibility — W3C Web Accessibility Initiative. Explores the rationale and organizational impact of digital accessibility standards.