Strategy & Industry
Content Management

Four Hundred Microsites and Nobody in Charge: Governing the University Web

A university-scale web inventory can easily turn up hundreds of distinct sites and microsites wearing the institution’s name: central servers, department servers, forgotten hosting accounts, and properties nobody has looked at in years.

Sites for initiatives that ended in 2006. Admissions microsites with outdated tuition figures still drawing search traffic. Templates from three brand generations coexisting like geological strata.

The crucial finding, and the reason this post exists: no one was negligent. Every site had a reasonable origin. A grant deliverable, a conference, a dean’s initiative. There was simply no system for anyone to be responsible to afterward.

Decentralization without governance does not produce chaos by malice. It produces it by default. Higher education’s structure makes it the default outcome unless someone designs against it. Here is the counter-design, from engagements where it has worked.

Why campuses grow four hundred sites

The diagnosis matters because it clears the people and indicts the system. Universities are federations. Schools, departments, centers, institutes. Each has real autonomy, separate budgets, and legitimate web needs that central services historically could not meet quickly.

So units solved their own problems. A departmental server here, an agency engagement there, a grad student’s weekend project that became critical infrastructure. Each solution was locally rational.

The sum is irrational for the institution. A web presence that leaks brand equity through inconsistency. Security exposure through unpatched sprawl, and our postmortems keep finding exactly these forgotten installs. Search authority lost through fragmentation. Visitor trust lost through the broken links and stale facts that fringe properties collect.

The fix is not centralization. Federated institutions reject transplanted org charts, correctly. The fix is federation with rules. Freedom inside a system, which is what governance means when it works.

The model, in five parts

  • The inventory, maintained. You cannot govern what you have not listed. The initial census is genuine work: crawling, DNS records, IT knowledge, and asking around. The professor’s server surfaces through asking around. Its maintenance is a process rule: no new site launches outside the registry. The registry’s columns do half the governing. Owner, purpose, platform. Last reviewed, sunset date.

  • Named ownership with review cycles. Every property gets a person. Not a department. A name. And a review pace matched to its stakes. The registry’s ‘last reviewed’ column turns abandonment from invisible to reportable. The annual review is where the 2006 initiative’s site finally gets its dignified retirement.

  • Standards with training attached. Editorial, brand, accessibility, and technical standards. Written once, taught continually. Standards without training are shelf documents. The schools that succeed treat editor onboarding the way they treat any credential, and the CMS permissions model enforces what training establishes.

  • Permissions that mirror actual authority. The multisite setups our platform coverage has tracked, WordPress networks, Drupal installs, exist precisely for this. Department freedom inside governed rails. Central security and brand control with local content control. One upgrade instead of four hundred. Platform consolidation is governance made structural. It is where the inventory’s long tail should migrate as reviews retire the ungovernable.

  • A decision forum empowered to say no. The piece most schools skip, and the one that decides whether the rest is real. A standing body with communications, IT, and academic seats, and actual authority. It approves new properties, settles disputes, and can decline a dean. If the forum cannot say no to power, the governance is decorative, and site number four hundred and one is already in procurement.

Making it survive contact with campus politics

The objections are predictable and answerable.

‘This is bureaucracy.’ The counter: the registry-and-review system is what lets requests move faster. A governed multisite platform can stand up a compliant department site in days, versus the months of ad-hoc buying the old chaos required. Governance done right is a service, and selling it as one is the whole political game.

‘Our unit is different.’ Sometimes true. The forum exists to hear it, and an exceptions process with reasons beats exceptions by default.

‘Who pays?’ The site’s current carrying cost usually funds the governed alternative with room to spare. The licenses, servers, and emergency fixes nobody totals. Run the total once and the CFO becomes governance’s best friend.

And the argument that actually moves cabinets, in our experience: risk. The unpatched microsite is a breach vector wearing the university’s name. The stale tuition figure is a lawsuit’s exhibit. The off-brand site is the institution’s front door for whoever lands on it.

Governance is how a federation keeps its promises. Every large spread-out organization, hospital systems, multi-division firms, has the same promises to keep. This is your diagnosis too. The campus just shows the problem most visibly. Four hundred sites visibly, in one case we could name but will not. The cabinet meeting recovered, the forum now exists, and the registry is down by ninety and counting. That path is available to anyone willing to count first.

A starting kit for the counting

Since audits begin somewhere. Your DNS zone files and certificate lists name more properties than anyone remembers. Your IT helpdesk’s ‘can you update the site’ tickets map the shadow site. A crawl from your own homepage finds what is linked. The gap between that crawl and the DNS list is precisely the orphan zone where risk lives.

Assign the census to whoever owns risk rather than whoever owns marketing. The framing determines the cooperation. And timebox it. A useful first inventory takes weeks, not quarters, and a partial registry governing something beats a perfect one governing nothing.

The four-hundred-site cabinet meeting is avoidable. The counting is not. Sequence accordingly.

And when the registry’s first draft lands on the cabinet table, lead with the two numbers that move rooms. How many properties carry the school’s name, and how many have an owner who knows it. The distance between those numbers is the governance case, stated in a way no one can unsee. Closing it, property by property, is the quietest high-value work in institutional web strategy this decade.