Brand system governance: how consistency survives growth

The central idea: A brand system scales when teams understand the…

A brand system scales when teams understand the principles behind choices and can request exceptions through a visible process.

Brand systems are usually launched as documents and judged as documents: how complete, how well-illustrated, how many components. That framing hides the actual problem, which is that a system is a behavioural contract between a central team and everyone who has to ship something on Friday. The document describes the contract. It does not enforce it, and it has no view on what happens when the contract is inconvenient.

What changed, and why it matters now: Track reuse, exception frequency, review time,…

The diagnostic is the shadow library. Somewhere in every organisation with a mature brand system there is a folder of assets that do not come from it — a deck template someone rebuilt, a colour that was needed for a campaign and never approved, a component copied and modified because the original did not fit. Its size is a direct measure of how hard the official route is. Teams do not build shadow assets out of rebellion; they build them because they had a deadline and the exception process had a two-week turnaround. The inventory is also unevenly distributed in a way that is worth reading. Divergence clusters in the teams furthest from the centre — regional offices, newly acquired units, agencies working to a brief — not because they care least but because they have the least access to the people who could approve something quickly. Governance that measures only headquarters output will report a healthy system while the edges run on their own conventions.

Build the operating model: Publish the ten decisions teams make most often, include…

Maintain tokens, components, voice rules, asset provenance, approval paths, and release notes in one source of truth. Treat exceptions as evidence for improving the system.

The exception route is the load-bearing part of the whole system and it is almost always the least designed. It needs a stated turnaround, a named decider, and an outcome that is not only yes or no — the third option, which matters most, is yes and this becomes part of the system. Without that third door, every approved exception is a permanent divergence, and the system accumulates undocumented special cases until nobody can say what it actually specifies. With it, the exceptions become the mechanism by which the system learns what it was missing.

Measure what the decision produced: A brand system is not a document that describes correct…

Track reuse, exception frequency, review time, accessibility defects, and the cost of correcting off-system production work.

Measure adoption at the artefact level rather than by survey. Count the proportion of shipped work that uses system components unmodified, lightly modified, or not at all, and read the middle number most carefully — heavy local modification usually means the component is close but wrong, which is more actionable than outright rejection. Time-to-exception-decision is the second measure, and it predicts the size of the shadow library about a quarter in advance.

Where execution breaks: A brand system scales when teams understand the…

Governance becomes either rigid policing or an unmaintained library. In both cases teams create parallel systems to keep shipping.

The dominant governance failure is a central team that optimises for correctness over throughput. It is an easy position to occupy, because every individual refusal is defensible, and the aggregate effect is that the system becomes something teams route around. The opposite failure is rarer and also real: a team that approves everything, at which point the system stops being a system and becomes a gallery of what people happened to make.

What this looks like in practice: Every brand system is consistent on the day it ships.…

What works in practice tends to look like a product team rather than a review board. A published response time for requests, a visible backlog, a changelog, and a contribution route that lets a team who solved a problem locally offer the solution back. The tone matters more than the process: a system whose owners ask what were you trying to do gets contributions, and one that asks why did you not follow the guidelines gets shadow assets and polite silence in meetings.

The strongest argument against this: A brand system scales when teams understand the…

The argument against this is that brand governance exists precisely to say no, and a system that optimises for convenience will drift into inconsistency one reasonable exception at a time. For organisations whose brand is their primary asset — luxury, regulated categories, anything where visual precision carries legal or price meaning — a slower, stricter route is not bureaucratic overhead but the point of the function.

That case is real and it is narrower than the people invoking it usually claim. The honest test is whether an inconsistency would be noticed by a customer or only by the brand team. Where the answer is the brand team, strictness is buying internal comfort at the cost of delivery speed. Where it is the customer, the strictness is earning its keep and the exception route should be slow on purpose, with that slowness stated rather than experienced as unresponsiveness.

A 30-day implementation sequence: Track reuse, exception frequency, review time,…

Publish the ten decisions teams make most often, include approved examples, and assign a response time for unclear cases.

Week one, audit the shadow library — collect what teams are actually shipping and compare it against the system. Week two, interview the five teams who diverged most and record why, without defending the system in the conversation. Week three, publish a response-time commitment for exception requests and staff it. Week four, close the three most common divergences by changing the system rather than by enforcing it, and publish the changelog so teams can see that requests produce change.

Track the shadow library, not the guidelines page

Keep a running inventory of non-system assets in active use, refreshed quarterly, with the team and the reason attached to each. It is an unflattering document and it is the only honest measure of governance health, because it counts what people do rather than what they agreed to. The inventory also makes the roadmap obvious: the three items appearing most often are the next three things the system should absorb, and that prioritisation is evidence-based in a way that internal wish-lists never are. Record alongside each item whether the team knew the system had a supported alternative. A surprising share of divergence is not disagreement but ignorance of what already exists, and that distinction changes the remedy completely: one calls for a new component, the other calls for better discovery. Systems that treat both as the same problem end up building components nobody needed while the findable ones stay unused.

Review adoption quarterly and the exception turnaround monthly. Read them together — adoption falling while turnaround lengthens is a single problem with two symptoms, and treating them separately produces a communications campaign when what was needed was staffing. When adoption rises, check whether the shadow library shrank; if it did not, the system gained new work rather than winning back existing work.

Editorial conclusion: A brand system is not a document that describes correct…

A brand system is not a document that describes correct output. It is a service whose customers are the teams shipping under deadline, and like any service it is judged on responsiveness rather than on the quality of its documentation. The organisations whose brands still look coherent at scale are rarely the ones with the strictest rules. They are the ones where asking was faster than working around. That property is not a cultural achievement; it is a staffing decision that someone made and kept funding, and it shows up in the response-time commitment long before it shows up in the work.