Why consistency needs an audit, not a reminder
Recognition is built by repetition. The distinctive brand assets framework reports the Ehrenberg-Bass Institute's argument that a cue only does its job when it is both widely recognised and uniquely linked to one brand, and the brand strategy guide adds that such cues are built through consistent use over years. Every execution that departs from the committed colour, mark or voice spends money on exposures that do not add to the same memory. That is the commercial case for consistency, and it is why the identity guide asks a named owner to run an annual audit of everything in live use.
A reminder email does not fix drift, because drift is rarely caused by people ignoring the rules. It is caused by structure: a template that was never updated, a retired logo still in the shared drive, a printer holding artwork from three refreshes ago, a location that was never given a template at all, or a guideline that is silent on the situation someone faced. The brand foundations hub describes the pattern: each departure is reasonable on its own, and the remedy is better templates, a named owner and an audit, not stricter policing. The audit is how the owner finds the structural causes.
It helps to be clear about what this audit is not. It does not test whether the strategy is right, which is the strategy guide's job, and it does not measure how famous or unique each cue is, which needs the survey described in the distinctive assets framework. It compares live materials with agreed rules. Social profiles can be scored in more depth with the social media presence audit scorecard, which this audit treats as one input rather than repeating. How the press describes the brand is a separate question, which the tutorial on running a PR audit answers from past coverage and conversations with journalists.
Before you start
Gather the reference documents: the one-page brand strategy, the current brand guidelines, the asset library and a list of the templates that exist. If the guidelines are missing or out of date, write the short version first with the brand guidelines template; an audit against unwritten rules turns into an argument about taste. Agree who owns the audit, who owns each group of touchpoints and who will decide disputed cases.
Set up one shared sheet before collecting anything, with the columns described after the steps. Decide how you will sample: every item for small groups of touchpoints, and a recent, representative selection where there are hundreds of near-identical items, such as product pages or local posters. Finally, agree the pass condition in writing: an item passes when it follows the guideline as written, not when the reviewer likes it. Where the guideline is silent, the item is logged as a gap in the guideline rather than a failure by its maker.
The steps
Set the scope and the reference documents
Decide which brands, markets and locations the audit covers and which version of the guidelines it measures against. A business with more than one brand audits each against its own rules, using the architecture map to decide which name belongs where. Write the scope at the top of the sheet so that nobody later argues an item was out of bounds.
List every touchpoint and its owner
Work through the places a customer, candidate, supplier or partner meets the brand: the website and app, ads in each channel, social profiles and posts, emails and automated messages, sales decks and proposals, invoices and order confirmations, packaging, signage, vehicles, uniforms, events, job ads, partner and directory listings. Put a named owner against each group. The owner is the person who can find the real materials and change them. In a business with branches, the replies each site posts under its reviews are a touchpoint too, checked against the reply guideline described in the guide to managing reviews across many locations.
Collect what is really in use, not the approved versions
Ask for live examples: screenshots of what customers see today, photographs of signs and printed items, the last few emails as received, the deck a salesperson actually used last week, the artwork a supplier holds on file. The approved file in the library tells you what should exist; the live example tells you what does. Gaps between the two are often the most useful findings.
Record each item against the guideline in one sheet
Give each item a row: where it lives, its owner, when it was last changed and a link or photo. Then check it against each element of the guidelines in turn, logo, colour, type, imagery, layout and voice, marking each as following the rule, departing from it, or not covered by the guideline. Check colour pairs used for text against the approved pairs, which the brand guidelines template says should pass the contrast requirements of the Web Content Accessibility Guidelines; the landing page accessibility guide explains why those decisions reach every page.
Audit the voice as carefully as the logo
The identity guide warns that logos are policed and words are not. Sample written material from every team that writes to customers: ads, onboarding and service emails, support replies, product descriptions, job ads and proposals. Read each against the brand voiceThe consistent personality a brand expresses in its words across channels, with tone adjusting to the situation and the audience. principles and their examples, and the vocabulary list, noting where the brand sounds like a different company. Record the situation, because tone rules by situation are often where the guideline is silent.
Rate each gap by how many people see it
Not every gap matters equally. Rate each one on two simple scales: exposure, from items seen by almost everyone the brand reaches to items seen by a handful of people, and severity, from a departure that breaks a committed asset, such as the wrong colour or a retired logo, to a minor slip such as an outdated spacing rule. A wrong colour on every delivery van outranks an old letterhead in one office.
Sort every item into keep, fix or retire
Keep what follows the guideline. Fix what is still needed but departs from it. Retire what should no longer exist: superseded templates, old logo files, materials for products or promises that have been dropped. Retiring includes removing the files from the asset library and shared drives, because the identity guide's warning holds: if a retired file can still be downloaded, it will be used. If the files are scattered across drives with no single approved copy, the tutorial on setting up a marketing asset library gathers them into one place with status, version, rights and expiry recorded.
Trace each gap to its cause
For every fix, ask why the item departed from the guideline. The usual causes are a wrong or missing template, an old file in a shared location, a supplier or agency working from old artwork, a team or location making its own materials, or a guideline that did not cover the case. Group the gaps by cause. In a business with many locations, the balance of central and local control in the guide to centralised and local marketing control usually explains a large share of them. Where a market is served by a distributor abroad, the gaps usually trace to the kit it was given, which the guide to marketing through a local distributor fixes before asking the distributor to replace what is live.
Assign owners and dates, fixing templates first
Fix causes before items. A corrected template prevents every future departure; a corrected poster prevents one. Rebuild or create the templates behind the largest groups of gaps, update the asset library, send suppliers the current files, and add any missing rules to the guidelines. Route new and changed templates through the creative approval SOP, and fold the asset library work into the naming and versioning routine in the creative production playbook. Then fix the remaining live items in order of exposure and severity, each with an owner and a date.
Re-audit on a cycle and report the trend
Repeat the audit on the cycle the identity guide recommends, with a short check of the highest-exposure touchpoints in between. Report one simple KPIA metric tied directly to an objective.: the share of sampled items that follow the guideline, by touchpoint group, compared with the last audit. The trend matters more than the level. A rising share shows the templates are working; a group that does not improve shows a cause that has not been fixed.
The sheet does most of the work, so it is worth setting up carefully. These columns cover what the steps need.
| Column | What to record | Why it matters |
|---|---|---|
| Touchpoint and location | Where the item lives and where people see it | Lets an owner find and change it |
| Owner | The person who can change the item | Fixes need a name, not a team |
| Evidence | Screenshot, photo or link to the live version, with the date | Shows what is in use, not what was approved |
| Element checks | Logo, colour, type, imagery, layout and voice, each marked follows, departs or not covered | Separates maker errors from guideline gaps |
| Exposure | How many people see the item, on a simple scale | Puts widely seen items first |
| Severity | Whether a committed asset is broken or a minor rule is missed | Protects the cues that build recognition |
| Decision | Keep, fix or retire | Turns findings into actions |
| Cause | Template, old file, supplier, local production or guideline gap | Lets one fix close many gaps |
| Action, owner and date | What will change, who will do it and by when | Makes the audit end in work done |
A garden centre group audits its brand (illustrative)
Tumbrock Garden Centres audits its brand
Tumbrock Garden Centres, an invented group of 7 garden centres, runs its first consistency audit before the spring season. The marketing manager lists 14 touchpoint groups and collects 212 live items over 3 weeks: signs, plant labels, loyalty emails, paid social ads, café menus, staff-made social posts and the delivery van livery. Against a guideline that specifies 1 green and 2 logo versions, the sheet records 5 greens and 3 logo versions in use, and 74 items, about 35% of the sample, departing from at least one rule. The voice check finds loyalty emails written in the brand's warm, plain voice but support replies that read like a utility company. Tracing causes shows that 41 of the 74 gaps come from 2 sources: an old shared drive that still holds a retired logo and colour file, and posters made locally because no poster template exists. The team deletes the old drive, builds a poster template with the headline, offer and photo editable and everything else locked, sends the sign printer the current files and adds a complaint reply example to the tone rules. It fixes the van livery and the plant label artwork next because almost every customer sees them, and leaves a dated letterhead in one office until the next reprint. A check of the highest-exposure groups 4 months later finds 88% of sampled items following the guideline, up from 65%.
If the audit shows that the identity itself no longer works, rather than that people are departing from it, the question changes from consistency to whether to refresh or replace the identity, and a replacement is put live with the tutorial on rolling out a rebrand. The brand foundations section places the audit alongside the strategy, identity and architecture work it protects.
Common mistakes
- Auditing the approved files instead of the live ones. The library shows what should exist; customers see what does. Collect what is really in use.
- Policing the logo and ignoring the words. Voice drifts silently across emails, support and sales unless it is sampled with the same care.
- Treating every gap as equal. A wrong colour on something everyone sees matters more than a slip on something few people see. Rate exposure and severity.
- Fixing items instead of causes. A corrected poster helps once; a template, a cleaned library or a supplier briefing prevents the next hundred departures.
- Leaving retired files downloadable. If an old logo can still be found, it will be used.
- Blaming makers for guideline gaps. Where the guideline is silent, the fix is a new rule or example, not a reprimand.
- Running the audit once. Without a repeat on a cycle and a reported trend, drift returns quietly.
- Auditing taste. The pass condition is the written guideline. Disputes about whether the guideline is right belong to a separate review of the identity.
