How to Run a Brand Audit: A Visual Designer's Framework
A five-part brand audit framework for visual designers: collect every touchpoint, sort by job, score consistency, trace the cause of drift, and report what to keep.

A brand audit is the step that decides whether the next project is a refinement, a system, or a rebrand. Skip it and the brief usually defaults to "make it look modern", which is the most expensive way to discover that half the brand was working.
I run audits before most identity and website work. At Datamotive the audit showed that three products, EasyMigrate, EasyHybridDR and EasyProtect, had grown up separately: different type weights on adjacent pages, decks on incompatible grids, and diagrams with arbitrary colour and line weight. At BLANKSY the audit covered the existing store and twenty competing sites before any identity work began. Neither audit was a slide of mood images. Both were evidence the team could argue with.
This is the brand audit framework I use. It is written for visual designers, but founders and marketing leads can run most of it themselves.
1. Collect every touchpoint, not the best ones
Ask for everything the brand has published in the last year or two: website pages, sales decks, social posts, ads, packaging, email templates, product screens, event graphics, signage, PDFs. Include the things nobody is proud of. The awkward partner one-pager made in a hurry tells you more about how the brand behaves than the launch campaign does.
Capture each item at the size people meet it. A homepage at 1440 pixels and at 375 pixels are two different touchpoints. A carton photographed flat and a carton on a shelf are two different touchpoints.
Name files with the channel, the date and the owner. That small habit pays off later, when you need to show which team or tool produced the drift.
2. Sort by job before you judge the look
Group the material by what it has to do, not by format. Typical jobs:
- Recognise: logo, favicon, packaging front, social avatar.
- Explain: product pages, diagrams, detailers, onboarding.
- Persuade: campaigns, ads, landing pages, sales decks.
- Support: documentation, invoices, help content, internal templates.
Sorting by job stops the audit becoming a beauty contest. A dense technical diagram and a campaign banner should not look the same. They should look related, with differences that come from their jobs rather than from whoever made them.
3. Score consistency on a short list of anchors
Pick the five to seven elements that carry recognition and check each touchpoint against them. My usual list:
- Logo usage: version, size, clear space, background.
- Colour: palette, proportion, and whether accent colours are doing structural work.
- Typography: families, weights, size ramp, line length.
- Layout: grid, margins, spacing rhythm.
- Imagery: photography style, illustration, diagrams, icon style.
- Language: tone, claims, calls to action, naming.
- Interaction: buttons, links, form patterns, motion.
Use three marks per anchor: consistent, drifting, broken. A spreadsheet works. What matters is that every judgement points to an actual file, so nobody has to trust your taste.
Language belongs on the list even in a visual audit. At Datamotive the visual system came together quickly, but the brand language took three rounds. A headline promising zero downtime contradicts a calm, exact visual identity, and buyers notice the contradiction before they notice the kerning.
4. Trace why the drift happens
Inconsistency is a symptom. The useful part of an audit is the cause. Common ones:
- No source of truth. Colours copied from screenshots, logos pulled from old decks.
- Missing assets. No compact logo, so people squash the full lockup into a favicon.
- Templates that do not fit real content. The headline box holds six words; the product team needs eleven.
- Several makers, no owner. Agencies, freelancers and internal teams each solving the same problem differently.
- Tools with their own defaults. Slide software, email builders and ecommerce themes all push their own type and spacing.
Write the cause beside each finding. "Product pages use four heading sizes" is an observation. "Product pages use four heading sizes because each was built from a different theme section" is a finding someone can fix.
5. Decide what the brand must keep
End the audit with three lists:
- Keep: assets that already carry recognition. At Datamotive the hexagon, the cube and the teal all stayed; the job was to strengthen them, not replace them. I wrote more about that decision in brand evolution vs rebrand.
- Fix: elements that are right in intent but inconsistent in execution, such as type weights, spacing or diagram line weights.
- Replace: elements that work against the brand's job, or that nobody can produce correctly.
This is also the moment to judge the scale of the next project. If most findings sit under Fix, the brand needs a system, not a new identity. The difference between documenting decisions and building a working system is covered in brand guidelines vs brand system.
Add the market view
An internal audit tells you whether the brand is consistent. It does not tell you whether it is distinct. For that, run a separate pass on competitors and adjacent categories. I keep it as its own exercise, described in competitive visual audit, because mixing the two makes it tempting to copy whatever the category leader does.
If the audit includes a website, pair this framework with a technical pass on navigation, performance and accessibility. My website audit checklist covers that side.
What the deliverable should look like
A good brand audit report is short enough to read in one sitting:
- One page on the brand's job and its audiences.
- The consistency scorecard, with links to the files.
- The causes of drift, ranked by how many touchpoints they affect.
- Keep, fix and replace lists.
- A recommended scope for the next phase, with what it will not cover.
Avoid long galleries of "before" screens with red circles. They feel thorough and get skimmed. A ranked list of causes gets acted on.
When to audit again
Re-run a lighter version every six months, or after any major launch. The second audit is faster because you are only checking the anchors and the known causes. If drift returns in the same places, the problem is governance, not design. Design system governance for small teams covers how to keep the fixes in place.
You can see how an audit turned into a single system in the Datamotive case study, and how a store and market audit shaped an identity in BLANKSY.