The 90-minute brand audit: our scoring sheet, free to steal

Every discovery we run starts the same way: before we talk about where the brand should go, we spend ninety minutes finding out where it actually is. Not a stakeholder workshop. Not a mood board. Ninety minutes of structured looking, scored against a sheet we have refined across a decade of brand engagements.

We give the sheet to clients at the end of every discovery, and now we're giving it to you. It is deliberately small — ninety minutes, four passes, eight criteria — because an audit that fits in an afternoon gets done, and an audit that gets done beats a perfect one that doesn't. Run it before a rebrand, run it quarterly as hygiene, run it when a new head of marketing asks "so how's our brand?" and you want a better answer than a shrug.

What this audit is (and isn't)

The 90-minute audit is triage. It answers one question: where is the brand leaking? It does not answer whether the strategy is right, whether the positioning is differentiated, or what the brand should become — that's the work of brand strategy, and it takes weeks, not minutes. Think of this as the blood pressure reading before the physical: fast, comparable, and an excellent predictor of where the deeper problems live.

The audit finds three kinds of leak, and the distinction matters because the fixes differ wildly:

Drift — the brand has eroded through a hundred small unmanned decisions. Fix: enforcement and better tooling.
Debt — the brand was never finished; assets are missing, so people improvise. Fix: build the missing pieces.
Mismatch — the brand as-built no longer matches the company as-it-is. Fix: that's a strategy conversation, and now you have evidence to open it.

Before the clock starts

Ten minutes of prep, then start the timer. You need:

One browser, twelve tabs. Homepage, pricing page, about page, one blog post, one product screenshot, the sign-up flow, a transactional email, a social profile, a slide deck, a job ad, an invoice or receipt, and the 404 page. These twelve are the brand as your audience actually meets it — which is to say, mostly not the homepage.
One reader. The single most valuable step: read the copy aloud. Voice problems are invisible on the screen and undeniable in the ear.
One scorer with no skin in the game. Whoever designed the homepage cannot score the homepage. Grab someone from a different team. Their squint is the instrument.

Pass 1 — The consistency sweep (25 minutes)

Click through the twelve tabs in order, fast, and write down every place the brand contradicts itself. You are hunting for the usual suspects:

Type drift. The site in one typeface, the app in another, the deck in a third. Count them. More than two type families visible across twelve touchpoints is drift; three or more is a design system that exists only as a PDF.
Colour drift. Screenshots the blues. They will not match. Brands that nail everything else routinely ship four distinct "brand blues."
Logo drift. Old mark on the social avatar, new mark on the site, a stretched mark in the deck footer. Logo drift is the least important of the three and the one everyone notices first — fix it last, but fix it.
Photography and illustration drift. Product UI screenshots next to stock photography next to three illustration styles from three eras. Each era marks a quarter where someone was in a hurry.

Score as you go. The sweep is deliberately quick because inconsistency is most visible at speed — this is how a customer's eye actually meets the brand, grazing across contexts, not studying one page in a boardroom.

Pass 2 — The voice sampling (25 minutes)

Now read. Take five samples: homepage hero, a product feature description, an error message or empty state, a sales email, and the copy on the pricing page. Read each aloud and ask three questions:

Could a competitor have written this? Paste the hero copy over a competitor's logo in your head. If it still works, you have wallpaper, not voice. Voice chart territory: the fix is teachable, but first you have to see it.
Is the register stable? Friendly homepage, legalistic errors, breathless sales email — a brand that changes personality by department reads as a company that doesn't know itself. Users notice. They can't articulate it, but they trust less.
Does the smallest copy match the biggest? Error messages and receipts are the honest voice of a company — nobody performs there. If your 404 page is charming but your cancellation email is a hostage note, the charming one is the lie. Score to the worst sample, not the best.

Pass 3 — The asset inventory (25 minutes)

Open the shared drive, the DAM, the "Brand assets FINAL v7" folder. You are not judging quality; you are checking completeness and findability against what the organisation actually needs each month:

Is there a logo file a non-designer can find in under two minutes, in the formats a non-designer needs (transparent PNG, not just an EPS)?
Do social templates exist for the posts the team actually ships — or does someone rebuild graphics from scratch every Thursday?
Is there a one-page brand sheet that sales can send without asking design?
Do the guidelines reflect what shipped this year, or what shipped at launch? (If the guidelines are a PDF, we have bad news and an article about it.)

The finding here is almost never "the assets are ugly." It's "the assets are unfindable, so people improvise, and the improvisation is what Pass 1 found." Asset debt creates consistency drift downstream. Fix the shelf and the mess stops regenerating.

Pass 4 — The mismatch check (15 minutes)

Last pass, and the only one that needs judgement rather than observation. Hold the brand next to the company as it is today and ask: does this brand still describe us? Signs of mismatch: the product has moved upmarket but the brand still looks scrappy-startup; the company now sells to enterprises but the copy still says "hey there!"; the flagship feature of 2026 isn't visually imaginable within the brand of 2022.

Score honestly on this one — it's the pass everyone inflates because the alternative is admitting the brand needs real work. That's fine. That is the finding.

The scoring sheet

Eight criteria, 0–2 each, sixteen points total. Score 0 for "failing," 1 for "uneven," 2 for "solid." Be stingy — a generous audit is a useless one.

#	Criterion	The 2-point standard
1	Type consistency	Two families max, everywhere, no exceptions
2	Colour consistency	Screenshots match; hex values survive contact with reality
3	Logo discipline	One current mark, correct usage, no orphans
4	Imagery coherence	One photography/illustration language across touchpoints
5	Voice distinctiveness	Copy no competitor could have written
6	Voice stability	Same personality from hero to error message
7	Asset completeness	Everything a busy team needs monthly exists and is findable
8	Brand–company fit	The brand describes the company as it is, not as it was

Reading the total: 13–16 and your brand is in genuine health — protect it, audit quarterly. 9–12 is drift territory: enforcement and tooling fixes, no rebrand, back in six months. 5–8 is debt: the brand was never fully built, and consistency work will fail until the missing pieces exist. Under 5 — or a mismatch score of 0 regardless of total — means the conversation isn't about consistency at all. Something structural needs to change, and you now have a scored sheet that says so, which is a much better way to open that conversation than a feeling.

What the patterns usually mean

After a decade of these, the score patterns repeat. High consistency with low distinctiveness (1, 2, 3, 4 strong; 5, 6 weak) is the competent-but-forgettable brand — common in B2B, fixable with voice work rather than visual upheaval. Low scores on 1–4 with a strong 8 often means a fast-growing company whose brand can't keep up with shipping — tooling and living guidelines, not a rebrand. And the pattern that predicts a real engagement: everything scores fine except 7 and 8 — the brand is well-kept and wrong, which is how you know the strategy, not the execution, is the problem.

If you want the picture in numbers you can trend over time, pair this with how we measure brand health — the audit is the qualitative snapshot, that piece is the quantitative baseline.

After the audit: three outputs, no deck

The audit's value collapses if it becomes a 40-slide presentation. Produce three things and stop:

The fix list — every specific defect from Passes 1–3, each with an owner and a two-week deadline. Most consistency fixes are days of work, not months.
The kill list — the assets, templates and touchpoints so far gone they're deleted rather than repaired. Deletion is the most underrated brand maintenance move there is.
One board slide — the scoring sheet and the total. Nothing persuades a leadership team like watching their brand get a number, and nothing reassures them next quarter like watching it move.

If the number says the problem is structural — or you'd like a second pair of eyes on whether it is — that's exactly the conversation we have with clients before anyone commits to anything as expensive as a rebrand.

Key takeaways

Ninety minutes, twelve touchpoints, eight criteria. An audit that fits in an afternoon is an audit that happens.
Distinguish drift (enforce), debt (build) and mismatch (strategise). The fixes are different; the confusion between them wastes quarters.
Read the copy aloud and score to the worst sample. The receipt and the error message are the honest voice.
Asset debt creates visual drift downstream — fix findability before enforcing consistency.
Score 0–2, be stingy, and let the pattern tell you whether you need tooling, voice work or a strategy conversation.
Output a fix list, a kill list and one slide. Never a deck.

FAQ

How often should we run this? Quarterly as hygiene if you're shipping fast, twice a year otherwise. The audit is calibrated for 90 minutes precisely so it survives contact with real calendars. If it grows, it dies.

Who should run it — in-house or an agency? In-house first, always. The exercise teaches the team to see the brand, which is worth more than our scores. Bring in outside eyes when the score says "structural" or when internal politics mean nobody can honestly score criterion 8.

Our guidelines exist but nobody follows them. What score is that? That's a 7 problem — asset completeness and usability — wearing a 1–4 costume. Guidelines nobody follows are usually findable-two-clicks-too-deep or styled for the launch case instead of the Tuesday case (see our piece on guidelines that get used).

Can we add criteria — accessibility, motion, localisation? Yes, but resist until you've run the base eight twice. Every criterion you add costs minutes, and the audit's power is that it's cheap enough to stay alive. Extend it only once the habit is unkillable.