Two competitors in the same category choose the same content management system, hire comparable partners, and spend similar budgets. Three years later, one has a website customers move through without help. The other requires an employee to explain it. The platform performed as specified in both cases. It was specified around assumptions about how a business makes, approves, and organizes what it publishes, and only one of the two companies worked that way.
The symptoms of structural misfit are common enough to pass as normal: templates duplicated instead of reused, a taxonomy someone maintains by hand because the products require content depth that was not initially planned for, an approval trail that lives in email while the system records only who published last. These are usually treated as content issues. But their roots often go back to decisions made years earlier, when the platform and its structure were chosen.
ACCESS IS NO LONGER THE ADVANTAGE
In the last few decades, technical capability was difficult to acquire, and assembling it early was worth paying for. Today, content platforms, search, personalization, analytics, and artificial intelligence-generated content are available to anyone willing to deploy them. Making these capabilities work well is still demanding. Having access to them no longer distinguishes one company from another.
The result is convergence. Competitors using the same platforms and methods arrive at the same language, navigation, page types, and category structures. The platform proposes a site structure, and the fastest build accepts it. The remaining advantage comes from the coherence of the decisions a company makes about how those capabilities should be used. That is design work in the structural sense: deciding what a system should make easy, what it should make someone stop and think about, and what it should refuse to do.
WHAT THE PLATFORM DECIDES FOR YOU
Every platform carries assumptions about the business that will run on it. Those assumptions live in the content model—the rules that define what a piece of content is, how content relates, what a customer record contains, and what a page is allowed to express. Those assumptions are reasonable because they were created for a general market, without knowledge of the company adopting them.
In one enterprise engagement, drawn from a pattern that recurs, a manufacturer with three product lines sells configurations that draw parts from all three. The default implementation assumes one product equals one page. To meet the deadline, the vendor generates a page for every catalog item. The resulting navigation asks customers to know which product they belong to before showing anything, even though identifying the right solution is precisely the help they need.
Technology and procurement chose the platform. The vendor configured the structure. The consequences land on customers, on the marketing team maintaining thousands of near-identical pages, and on salespeople answering by phone what the navigation could not.
A leadership team sharpening its positioning finds that the terms in the draft came from a category list a developer created during migration. Those categories became section names. The section names became search terms. And the search terms became the words customers use in inbound inquiries. By the time anyone questions the vocabulary, no one can say when it was decided.
THE HARDEST DECISIONS GET THE LEAST ATTENTION
Technology commitments are difficult to reverse. Contracts run for years. Migration costs and retrained staff push reconsideration further out. Design decisions are understood to be changeable, so they get scheduled later and carry less authority.
The result is an inversion. Companies spend executive attention on typography, photography, and a homepage many customers will never enter through. Meanwhile, the content model that governs every page is settled by a solutions architect working from a template, in a conversation the brand and product teams do not attend.
Design sits under marketing. Platform engineering sits under technology. The two are joined by a project plan instead of a shared business model. A project plan can sequence work. It cannot settle a disagreement about what the company is.
WHAT THE NEW TOOLS MULTIPLY
AI-generated content puts this under pressure. Producing a page, a variant, or a product description is now inexpensive enough that output is no longer limited by how many people can write it. The constraint has moved from what a company can produce to what should exist.
Many organizations will leave that second question unasked and use the new capacity to reproduce the inherited model faster, since the capability arrives with no reason to reexamine the structure it fills. A company working from a model it authored absorbs the capability. A company without one scales the condition it was in.
HOW TO DECIDE BEFORE BUYING
Organizations that handle pre-purchasing decisions well make a few changes that require no new technology and resist retrofitting:
- Design leadership takes part in platform evaluation.
- The content model is treated as a strategy document that happens to be built in software.
- Selection criteria include what a platform makes difficult, since constraints shape behavior more reliably than features.
- Category structures are reviewed by the people accountable for how the company describes itself.
This sounds easy to support, but it changes how platform decisions get made. The requirements document is written by technology and procurement and scored on feature coverage and cost, with design arriving at the demo stage. Contesting that document before it goes out adds weeks and forces the business to say what it is before a vendor says it for them.
Three years after two companies buy the same platform, the difference between them is easy to see and cannot be explained by platform capability or implementation quality alone. One can still explain itself clearly after strategic change, product expansion, and leadership turnover. The other has spent the same years accumulating exceptions, each defensible when granted but collectively responsible for a digital property no one inside the company can describe in a sentence. The model was either authored or inherited long before anyone reviewed a design.
Goran Paun is the principal and creative director at ArtVersion.