How to Run a Brand Audit, Step by Step

Brand
Written By wevisualise designs

A brand audit is not a discussion about whether the logo still feels right. It is a count, a measurement and a comparison, and it produces a list of defects with owners and dates against them. This is the method we use, including the accessibility thresholds most audits skip and a worked example with real contrast figures.

What a brand audit is for

An audit answers three questions with evidence. What assets exist and where are they in use. Where does the actual application differ from the intended one. What is that gap costing in recognition, credibility or legal exposure.

Consistency is the commercial argument. Research by Ipsos and Jones Knowles Ritchie, reported by Marketing Week, tested more than 5,000 brand assets with over 26,000 consumers and found only 15% were "truly distinctive", with 65% falling into the lowest tier. Logos reached the top grade in 19% of cases and brand colours in just 4%. The research points to sustained, consistent use rather than initial design as the thing that makes an asset recognisable. An audit is how you find out whether yours is being used consistently enough to work.

Budget roughly two to five working days for a business with a single site and one website. More if you have vehicles, multiple premises or a reseller network.

Step one: build the asset inventory

Open a spreadsheet. One row per asset, with columns for asset name, file format, version, where it is stored, who owns it, where it is currently in use, and last updated.

Cover at minimum: logo files in every format and lock-up, colour values as hex, RGB, CMYK and Pantone, typefaces with licence details, iconography, photography library, templates for documents, presentations and email, social profile assets, signage artwork, vehicle livery, packaging and print artwork.

Count the duplicates. On most first audits the interesting number is how many versions of the primary logo are in live circulation, and how many of them are wrong. Anything you cannot locate a source file for goes on the list as a risk, because it will need recreating the first time someone asks for it at a different size.

Step two: review every touchpoint, not a sample

Sampling hides the worst offenders, because the worst offenders are usually the oldest and least visited. Work through the full list.

Digital: website by template type, not by page count, so homepage, service page, blog post, contact form, checkout, error page. Then email signatures, transactional email, invoices and quotes, proposal templates, every social profile including the ones nobody has posted to in two years, review site profiles, directory listings, job adverts and recruitment pages, and any partner or reseller site displaying your mark.

Physical: signage, vehicles, workwear, packaging, printed literature, exhibition materials, business cards, and anything laminated near a reception desk.

Your search listing counts too. Google Search Central states that "Google's generation of site names on the Google Search results page is completely automated and takes into account content from a site's home page and references to it that appear on the web", and that "WebSite structured data is most important, if you want to specify a preference". Google also notes that site names are supported at domain or subdomain level only, not at subdirectory level. If your business appears in search under a name nobody at the company would choose, that is an audit finding with a specific fix.

Score each touchpoint against three states: on brand, minor deviation, or off brand. Record the specific fault, not a judgement.

Step three: run the accessibility & contrast checks

This is the section most brand audits skip, and it is the one with measurable pass and fail lines rather than opinions.

The Web Content Accessibility Guidelines are a W3C Recommendation, with version 2.2 published on 12 December 2024. W3C defines three conformance levels: "A (lowest), AA, and AAA (highest)". AA is the working standard. In the UK it is also a legal standard for parts of the economy: GOV.UK states that public sector bodies must meet the "Web Content Accessibility Guidelines (WCAG) 2.2 AA accessibility standard" and publish an accessibility statement, under the Public Sector Bodies (Websites and Mobile Applications) (No. 2) Accessibility Regulations 2018. If you sell to councils, the NHS or central government, your assets sit inside their compliance boundary.

The thresholds, taken directly from the W3C Understanding documents:

Criterion Level Requirement
1.4.3 Contrast (Minimum), normal text AA "a contrast ratio of at least 4.5:1"
1.4.3 Contrast (Minimum), large text AA "at least 3:1"
1.4.11 Non-text Contrast AA 3:1 for user interface components and graphical objects
1.4.6 Contrast (Enhanced), normal text AAA "a contrast ratio of at least 7:1"
1.4.6 Contrast (Enhanced), large text AAA "at least 4.5:1"

Two details matter in practice. W3C defines large text as "at least 18 point or 14 point bold", noting that "14pt and 18pt are equivalent to approximately 18.5px and 24px". And W3C is explicit that these are thresholds, not targets to round towards: "the computed values should not be rounded (e.g. 2.999:1 would not meet the 3:1 threshold)". W3C also confirms that "text that is part of a logo or brand name has no contrast requirement", which is why a logo can pass while the button beside it fails.

A worked example with real numbers

Here is a real finding, from our own site. It is a fair example because it is the kind of defect that survives for years without anyone noticing, and because we would rather show a measured failure than a tidy hypothetical.

Every button on wevisualise.co.uk used a cream label on the orange highlight colour. Measured against the WCAG relative luminance formula, on the orange the site actually rendered:

Colour pair Measured ratio AA normal text AA large text AAA normal text
Orange with cream label, as it was 2.81:1 Fail Fail Fail
Orange with white label 3.46:1 Fail Pass Fail
Orange with Off Black label, as it is now 5.03:1 Pass Pass Fail
Off Black on white 17.44:1 Pass Pass Pass

The cream label failed at every size, including the large hero button, which is the harder failure to spot because large text only needs 3:1. Even white, the colour most people would reach for, would have failed for normal text. Off Black on the same orange measures 5.03:1 and passes.

The audit also surfaced a second defect worth naming: the orange rendering on the site was two points away from the orange documented in the guidelines. A brand audit finds that sort of drift, and a colour that is nearly right is a colour nobody can specify.

The fix cost nothing. Switch the button label to the Off Black already in the palette, and every call to action passes. Note what the audit did not recommend: changing the orange. The colour is a recognition asset. The text colour on top of it is not.

Run this check on every fixed colour pairing in your palette and record the number, not a verdict. Numbers survive arguments about taste.

Step four: test message consistency

Collect the first sentence of your description from ten places: homepage hero, About page, LinkedIn company page, Google Business Profile, the last proposal you sent, an email signature, a recruitment advert, a directory listing, a printed leaflet, and whatever the receptionist says when asked what the company does.

Put those ten sentences in one column and read them together. If a stranger could not tell they described the same business, you have a messaging problem rather than a design one.

While you are there, check the claims. The CAP Code, enforced by the Advertising Standards Authority, requires marketers to "hold documentary evidence to prove claims that consumers are likely to regard as objective and that are capable of objective substantiation" before publication. Any unevidenced claim on the list is a finding.

Step five: map the competitive frame

Take four to six named competitors, capture their homepage, primary claim, colour palette and typography, and date the capture. You are looking for two things: where you are indistinguishable from a competitor, and where a distinctive position sits unclaimed.

Colour is the usual failure. If three of six competitors use the same blue and you are the fourth, your palette is doing no identification work at all, which is consistent with the finding that only 4% of brand colours in the Ipsos and Jones Knowles Ritchie study reached the top distinctiveness grade.

Step six: gather customer perception evidence

The audit is incomplete until someone outside the business has spoken. You do not need a research budget. The GOV.UK Service Manual advises that for qualitative work such as interviews and usability testing, "you'll usually need between 4 and 8 participants for each round", and recommends aiming "to do at least one round of research every 2 weeks", with more small rounds preferred over one large one.

Six recent customers is enough for a first pass. Ask what they would say the company does if a colleague asked, what nearly stopped them buying, and which competitor they also considered. Then show your logo, palette and strapline with the name removed and ask what they identify. Anything correctly identified is an asset to protect. Anything not identified is an asset you are free to change.

Step seven: write the output document

The deliverable is not a presentation. It is a defect register with a summary on the front.

Each finding needs: what was found, where, evidence in the form of a screenshot or a measured figure, severity, the recommended fix, an owner, and a fix-by date. Severity should be scored consistently, for example: critical for legal or accessibility failures, high for anything a customer sees at the point of purchase, medium for internal templates, low for cosmetic drift.

Front the document with counts. Number of assets inventoried, number of touchpoints reviewed, percentage on brand, number of critical findings, and the three fixes that would close the most findings at once. Those five figures are what a leadership team acts on. Everything else is the appendix that proves them.

What to do next

  1. Start the inventory spreadsheet today and give it a single owner. Two hours will tell you how many versions of your logo are in circulation, which is usually the most persuasive number in the whole exercise.
  2. Measure your two most-used colour pairings against the 4.5:1 and 3:1 thresholds this week. If a call to action fails, it is often a text colour change rather than a palette change.
  3. Run the ten-sentence description test. It takes twenty minutes and needs no budget.
  4. Book six customer conversations for the next fortnight, following the four to eight participant guidance, and ask the recognition question with the name removed.
  5. Set the severity scale before you start finding things, so the register cannot be argued into a wish list.

An audit is only worth the fixes it triggers, which is why the output has to be a register rather than a critique. If you would rather have the measurement, the register and the fix plan produced independently of whoever built the assets, that is part of our brand identity & strategy work. The findings are usually cheaper to fix than clients expect.

Sources

Related reading

wevisualise designs https://wevisualise.co.uk
Previous

Naming & Straplines: Getting the Words Right
Next

What a Brand Strategy Actually Contains