Series note. Third and final piece on higher ed positioning. The first, How Universities Find A Brand Position They Can Actually Hold, was about how higher ed institutions can carve out a defensive position. The second, A University Is a City, was about where that position lives across the institution, its faculties, and its programs. This one is the operating manual: how to run the position day to day so it doesn’t get negotiated away once everyone starts using it.
tl;dr
Brand governance in higher education is the operating model that decides how a university’s positioning is used, adapted, and protected once it leaves the strategy document and spreads across faculties, programs, and campaigns. The workable version is centralized decentralization: a center that owns the logic, the standards, and the hard decisions, with faculties and units free to make the brand specific inside clear guardrails. It runs on a Lead and Support split, defined decision rights, proof standards, and a network of working hubs. Arguably it’s under more pressure now, because AI lets any team produce plausible content faster than a slow center can review it.
The instinct when a position starts slipping is to clamp down: tighten the guidelines, add an approval step, point everyone at a sternly worded brand portal. None of it works, because the problem was never that people lacked rules.
McKinsey has spent years making the general version of this argument: what turns a strategy into performance is the operating model beneath it, and most strategies die in the gap between the two. A brand position dies there too.
So treat this piece as an operating manual written for the person who has to run this: a Director of Brand who knows the current setup is too loose, that pulling every decision into the center would be slow and politically impossible, and has roughly none of the team the textbook answer assumes. What follows is the model we’d hand that person, and the tools inside it that can be standing up by next quarter.
It’s a real shame when a legitimately useful framework gets a cringy, forgettable name, and this is one of those times.
Total central control is a fantasy in a university, where a faculty can outlast three presidents and a donor can reorder priorities with one phone call. At the other extreme, total local freedom gives you fifteen schools that each sound like a separate institution, and a parent brand nobody recognizes. The workable answer is a hybrid, and for once the insufferable name fits: centralized decentralization.
The center owns the logic of the brand: the position, the hierarchy, the standards, the decision rights, the shared tools. The edges own the expression, because they have the audience knowledge, the disciplinary nuance, the local credibility, and the speed the center never will. The center doesn’t make every move; it just makes sure the moves still belong to the same institution. Everything else in this piece is the machinery that turns that idea into something that runs.
The first practical question is which work the center does itself and which it equips others to do. The cleanest version of this idea comes from outside higher ed entirely. In Team Topologies, Matthew Skelton and Manuel Pais describe a platform team whose job is to make every other team faster and to stop them rebuilding the same things from scratch, while staying out of the approval path. It was written for software teams, but the idea ports directly: the center exists to make everyone else faster without getting in the way. Most teams can sort the split in a single working session, and it pays for itself the first time a program manager stops emailing central for permission to do something they were always allowed to do.
The center leads where reputation, strategy, and risk are greatest: institutional campaigns, the parent-brand claim, the brand architecture, the high-stakes recruitment narrative, advancement storytelling, the core web templates, accessibility standards, and the proof requirements everyone else builds against. These are the calls where one bad decision shows up across the whole institution, so they don’t get delegated.
Support covers most of what a university publishes day to day: program pages, faculty profiles, event promotion, student and alumni updates, the work of turning research into something a human will read. The unit doing that work knows its audience in a way the center never can, so it makes the call, and the center keeps its hands off the approval button.
Lead and Support settles who does the work; decision rights settle who gets to decide, the question institutions dodge hardest, because pinning it down makes enemies while leaving it vague feels diplomatic, right up until you notice an unanswered decision right doesn’t disappear. It goes to whoever has the most seniority, urgency, or stamina, which is how exceptions quietly turn into policy without anyone admitting it.
Sort decisions into four kinds and most of the fights resolve themselves before they start.
Center decisions touch the whole institution: parent-brand language, the naming system, major campaign platforms, accessibility, proof standards. These can’t be renegotiated faculty by faculty without the position coming apart.
Local decisions belong to the unit, because the unit has the better context: which employer partnership to feature, how a specific cohort talks, which student work proves the point. The center has no business approving these and looks ridiculous trying.
A third kind needs consultation, because it crosses more than one part of the system: a faculty repositioning, a recruitment campaign, an advancement narrative that draws on an institutional claim.
The last kind needs escalation, because the institution is about to trade away something it said it wanted to own. Everyone designs around this category and assumes it won’t come up. It comes up constantly, and it gets its own section below.
Sorting decisions this way does something quiet but useful. People stop fighting over who’s right and start agreeing on where a decision belongs, and “this is a center call” goes down a lot easier than “your idea is worse than mine.”
A position closes the gap between what an institution claims and what it can show. Governance is what stops that gap from reopening, one borrowed claim at a time, and the rule is almost embarrassingly simple. A claim is allowed at the level where the evidence lives, and the bolder the claim, the more evidence it takes to make it.
An institutional claim has to be true in enough places that someone who goes looking keeps finding it. A faculty can sharpen that claim inside its own domain, where it has the proof. A program, in almost every case, should be proving the claim above it instead of inventing a new one, through the things a program can point to: the venture lab, the co-op data, the placement record, the named companies its graduates started.
A strong central claim is the best-sounding language in the building, so of course it spreads. A faculty that can’t prove it uses it anyway because it reads well, a program one level down copies that version, and a few levels further, the boldest line on the website is sitting over evidence a single follow-up question would expose. You can’t point at the person who broke it, because nobody did. Every borrow looked reasonable to whoever made it, which is what makes the drift so hard to stop. Proof standards are the brake, and they only work if someone is willing to say no when the proof isn’t there, even when the person asking outranks the brand team and wants an answer quickly.
Most governance models jump straight from central standards to local execution, and they break in the gap between the two. That gap is where someone in a program office is staring at a new message framework, unsure whether it even applies to them or whether their proof is strong enough to publish. If that middle layer has no structure, the institution learns slowly and unevenly, usually too late to matter.
A network of hubs gives the middle layer a shape. A hub is a small, recurring working group built around a real operating need rather than an org-chart box: recruitment content, program marketing, advancement storytelling, research communications, social, web publishing, analytics, accessibility, AI content. The people in it are the ones producing the work, plus someone from the center who can move decisions back and forth.
A hub is only worth the name if it works differently from a committee. A committee is where decisions go to be admired and deferred; a hub moves knowledge, sharing what’s working, surfacing friction early, training the people who’ll never read the full guidelines, and giving the center a live read on where the system is working and where it’s drifting, months before that drift shows up on the website.
In practice that looks like a recruitment content hub keeping program pages from sliding into the same interchangeable employability copy, a research communications hub helping institutes translate specialized work without detaching from the institutional position, and an AI content hub stopping a hundred plausible rewrites from flattening the university into the same warm sameness as everyone else. A working hub is the opposite of another standing meeting: a layer of people who can make good brand calls without escalating every one, which is the only thing that scales when the center is too small to police everything itself.
This stays theoretical until someone powerful wants an exception, which always happens, so here’s the scene.
A dean brings in a campaign for a new program with a major gift attached, the kind that names a building, and a donor who has opinions. The launch date is fixed because the announcement is tied to the gift, and somewhere in the creative the faculty’s territory has widened into language that belongs to the institution, the broad claim the parent brand keeps for itself. It’s a reasonable request, and it’s also the one that, repeated across forty deans, dissolves the position entirely.
The trouble is that the brand lead can only say the campaign feels off-strategy and the dean can only say their audience is different. Once it’s one set of instincts against another, a feeling gives power nothing concrete to push against, so the person holding the building-naming gift and the fixed launch date wins. Lean on taste and you’ve already handed the decision to whoever has the most leverage.
What gives you a fighting chance is a short set of filters the institution agreed to in advance, back when no specific gift was on the table and everyone could afford to be principled. A handful of blunt questions do most of the work: would this still read as our university with the logo removed? Is there evidence under the claim, or just confidence? Can a program use the language without distorting it? Will it still mean something when this president is gone and the next one inherits it? And does it pull against the one thing the institution said it would own? None of these settles the argument on its own, but together they move it off “I like it” and onto something the dean has to answer to, which is the only kind of argument a brand team ever wins against someone who outranks it.
And sometimes the honest answer to those questions is that the dean’s right: the territory was drawn too tight, the faculty has proof the center missed, and the exception should be granted and the standard updated to match. A governance system that can never say yes to a powerful person is just brittle, and brittle systems get worked around. The filters won’t win every argument, and they’re not meant to; they exist so that when the institution stretches its position, someone made that call on purpose and can defend it, instead of it slipping through as an accident nobody remembers approving.
All of this depends on one thing most models skip: none of these filters mean anything if the brand lead has no authority. In most universities, the person responsible for the brand outranks nobody who can break it. A director in central marketing can’t tell a dean no in any way that sticks, not unless the president or provost delegated that authority in advance and is willing to defend it when the dean appeals over their head. Arizona State’s inclusion-at-scale position didn’t last because the brand team guarded it well. It lasted because Michael Crow restructured the university around it and spent twenty years defending it. Decision filters without delegated authority are just a more articulate way to lose the argument.
The reason most of this fails has little to do with the model and almost everything to do with what happens after the kickoff. Universities love launches because launches feel finished. The platform’s approved, the campaign’s live, the guidelines are published, everyone moves on. Then the territories drift, the program pages calcify, a three-year-old campaign keeps running because no one owns the decision to kill it, and a claim that was aspirational in 2023 is provable now while the language still hedges.
So the work that protects a position is mostly maintenance, and a maintenance calendar is the most useful thing this whole model produces: scheduled reviews of the faculty territories, audits of the program pages, a proof library updated as the evidence changes, and a standing check on new campaigns, microsites, and AI-generated content before they drift. Written out, it sounds tediously operational, which is the point. The upkeep is the part that protects the position, long after the launch everyone celebrated is forgotten.
That upkeep needs a measure, and the measure can’t be output. A university that counts pages, posts, and campaigns will produce more pages, posts, and campaigns and look busy while getting harder to understand. Track whether the place got clearer instead: whether a prospective student can explain what makes you different, whether a donor understands the future they’re funding, whether a faculty recruit believes the claim. This is what CASE’s work on brand and reputation metrics is for. It tracks awareness, trust, reputation, and preference on a consistent basis, which is how you learn whether the position is getting through, instead of counting posts that only prove the team was busy.
Let’s be honest about the part that kills most of these systems: it’s all overhead. The hubs, the training, the proof library, the audits, the standing authority to enforce a decision, none of it looks good on a budget line, and overhead is the first thing a CFO trims in a soft year, which tends to be exactly the year the position is eroding fastest. The honest version is that you can’t have any of it for free. Funding the support is what keeps decentralized expression coherent, delegated authority is what makes enforcement real, and a sharp position only stays sharp if the institution can stomach disappointing people who outrank the brand team. Most failed brand systems are an expensive search for the version where none of that is true.
Every problem in this piece now has a multiplier on it. A faculty team can generate a full campaign in an afternoon, a department can rewrite its pages without grasping the hierarchy they’re meant to fit into, and a program can spin up twenty versions of a claim that are all fluent, all plausible, and all disconnected from anything the institution can prove.
The worry that AI will make university brands sound robotic has it backwards. The real risk is that it makes them sound professionally reasonable: warm, competent, agreeable, interchangeable, which is the exact kind of language higher ed already overproduces by hand. Weak governance plus generative AI is a machine for making more of it, on every page, faster than any center can read.
Pointed at approved claims, a real proof library, defined territories, and clear tone guidance, the same tools let a small central team extend its judgment instead of drowning in volume. That’s why AI guidance belongs inside brand governance rather than an IT policy nobody in marketing opens. The questions are practical and worth settling now, while the volume is still building: what source material a team can use, what claims need proof before anything ships, who reviews public-facing AI work, and what should never be generated at the program level at all.
Strip away the apparatus and governance is one capability: the institution’s ability to tell someone who outranks the brand no, for a reason everyone in the room can follow, and make it stick. The Lead and Support split, the decision rights, the proof standards, the hubs, the filters, the maintenance calendar, all of it exists to make that one moment winnable for the person who has to be in it.
A university rarely needs more language. It needs discipline around the language it has already chosen, and someone with the standing and the tools to defend it when defending it gets expensive. Find the position, organize it across the institution, then govern it. Skip that last part, and you watch it get negotiated back toward the average, one reasonable exception at a time, until even the people inside can’t say what the place is for.
Brand governance in higher education is the operating model that determines how a university’s positioning is used, adapted, protected, measured, and maintained across the institution. It defines who owns the institutional claim, what evidence a claim requires, which decisions are local, and what happens when a powerful unit wants an exception that pulls the institution away from the territory it chose to own.
Centralized decentralization is a hybrid governance model where the center owns the institutional position, hierarchy, proof standards, shared tools, measurement, and decision rights, while faculties and units adapt the brand for their own audiences within clear guardrails. It fits universities because authority is distributed across faculties, donors, and leadership, so pure central control is rarely real.
The Lead and Support model sorts brand work into two categories. The center leads where reputation and risk are highest: institutional campaigns, the parent-brand claim, brand architecture, accessibility, and proof standards. The center supports everywhere local knowledge wins, through templates, modular messaging, proof libraries, training, and clear workflows, so units can produce strong work without the center approving it line by line.
A brand hub is a small, recurring working group built around a real operating need, such as recruitment content, research communications, social, or AI content. Unlike a committee, a hub moves knowledge: it shares what works, surfaces friction early, trains local users, and gives the center a live read on where the brand system is working and where it is drifting.
Decision rights define who gets to decide what, sorted into four kinds. Center decisions affect the whole institution and cannot be renegotiated unit by unit. Local decisions belong to the unit with better context. Consultation decisions cross multiple parts of the system. Escalation decisions involve a trade-off against the position itself. Naming the category before the argument removes most of the politics.
AI raises the cost of weak governance by letting decentralized teams generate large volumes of plausible, generic content quickly. The risk is not that brands sound robotic, but that they sound professionally reasonable and interchangeable. With messaging hierarchy, approved claims, proof libraries, tone guidance, and review workflows, the same tools let a small central team extend its judgment instead of scaling undifferentiated language everywhere.
Brand strategy finds the position, the white space and cultural role a university can credibly hold. Brand hierarchy organizes it, giving the institution, faculties, and programs distinct messaging jobs and closing the proof-promise gap. Brand governance protects it through centralized decentralization, deciding who can stretch the position, what proof a claim needs, and what happens when a powerful unit wants an exception.
