In September 1878, after Benjamin Franklin left the first U.S. Constitutional Convention, he was asked what kind of government delegates had settled on. “A Republic,” he replied, “if you can keep it.” The implication: defining models and structures was only the beginning of an ongoing struggle.

The same is true of information architecture.

Often, we think of IA in the context of a major project such as a website design or redesign. That is, a big one-off lift towards sound information structures. It’s like a constitutional convention: a rare event that requires great focus and alignment. But what happens after launch?

If the system is to remain relevant, it must change. Over time, inconsistencies and exceptions creep in. Structures that were initially clear become muddled. Like other systems, IA needs upkeep. But changes can’t be arbitrary. Structures must adapt to changing conditions without sacrificing clarity, coherence, and cohesiveness.

That’s why governance matters. Defining clear policies and methods upfront eases the job of maintaining an information architecture. Alas, IA governance docs are relatively rare. I’m currently helping a client craft a set of IA governance policies, and thought it a good time to review some pointers for good IA stewardship.

1. Establish clear boundaries

Few people in orgs understand what information is, much less how to govern it. It’s especially tricky because there’s overlap with other disciplines. Content strategy in particular is difficult to distinguish from IA.

For example, managing taxonomies is clearly within IA’s remit, but folks often face taxonomy-related challenges while editing content. The need for a new content category often arises when publishing a new content.

There are also different roles within the org, and they’re all likely to have different takes on site-wide structures. Clarity around who’s responsible for what goes a long way to making the process of managing the IA easier.

Someone must be the steward of the information architecture. They might be responsible for other content tasks, but it’s important that they understand where IA ends and content begins — without becoming doctrinaire about it.

It helps to keep the ultimate goal in mind: keeping information in the system findable and understandable. The steward must have the tools and wherewithal to help others in the org how their choices might affect findability and understandability, and why that matters to the business.

What you can do:

  • Clearly define what you’ll manage. In the case of IA, it’s structure, not content. This includes site-wide constructs such as taxonomies and navigation structures.
  • Define clear roles and responsibilities. Understand when you’re wearing a content manager hat and when you’re wearing an IA hat. Understand which stakeholders influence information structures on the site, what matters to them, and how to best communicate.
  • Adjust rules to suit new needs. Rules and guidelines are meant to help, not hinder. Use common sense and opt for practical solutions. Emerson said, “A foolish consistency is the hobgoblin of little minds, adored by little statesmen and philosophers and divines.” You are neither; act wisely.

2. Enable coherent evolution at scale

I realize this sounds a bit academic. The point is that the site’s structures should change without losing their character. This requires understanding what that character is supposed to be. What kind of place is your site? What do people expect to do and be there? Are there sub-sections of the site with different needs? Etc.

The key is making it possible for the architecture to adapt without sacrificing clarity and coherence. For example, you should be able to add items to a navigation menu without unintentionally changing how users understand the context created by the set of labels used in the Navy structure.

This requires understanding the bigger picture. Individual nav items (or terms in a taxonomy) don’t function individually; they’re always part of a broader set. That set creates a context that changes how users interpret the meaning of individual terms. Content and context are tightly coupled and interdependent.

What you can do:

  • Keep the big picture in sight. It’s easy to get lost in the weeds. While it’s important to review terms in taxonomies, it’s also important to review higher-level artifacts such as conceptual models.
  • Manage contexts as well as content. Sets of terms create contexts in users’ minds. Consider changes from multiple perspectives: tweaking a couple of terms might completely change the way users understand the context.
  • Establish periodic IA reviews. While it’s important to deal with granular changes as they come up, it’s also important to take a step back every once in a while to see how structures are performing overall. For many orgs, a yearly cadence should be enough.

3. Keep it relevant

Governance won’t work if it’s relegated to a Google Doc nobody visits. While it’s useful to document policies, the point is implementing them. This requires that governance guidelines be followed, not merely documented.

Thus, guidelines and policies must be usable and actionable by the people who need them — namely, site and content managers. This implies particular forms for documentation, processes to induct new folks into the system, and guidelines for keeping things guidelines themselves relevant.

Which is to say, IA governance docs need IA (and governance) too. Implement actionable feedback loops throughout the system, including in the guidelines themselves.

What you can do:

  • Write guidelines with future collaborators in mind. While it may feel wasteful, provide a bit more context than what you and your current team needs. Make it easy for other folks to pick up where you leave off.
  • Keep documentation separate from implementation. Don’t work directly in production. Hash out IA changes in shared spreadsheets and other such indirect documents and not in the site’s content management system.
  • Clarify feedback mechanisms. Understand the inputs you’ll use to determine whether the IA is serving user needs. Provide means for other team members and stakeholders to suggest changes at different levels in the system.

    Keep folks in the loop

As you may have noticed, many of these recommendations imply clear communications. Keeping the IA relevant isn’t a solo enterprise.

Establish regular review cycles with stakeholders from different parts of the org. Check in with folks. Ask how their needs are evolving. Share changes you’re implementing elsewhere. Ensure the IA evolves in alignment with business needs and user expectations. Bring others in to distribute the load of upkeep.

IA isn’t just about structuring stuff to make it more findable; it’s also a manifestation of what the organization aspires to be in the world. That is, the IA should represent a clear, tangible articulation of the org’s strategy. As such, information architecture governance isn’t only about keeping the IA relevant: it’s also an opportunity for the org to check in with itself to ensure it’s bringing its best self forward.

Because of this, IA stewards have a critical role to play. While it’s work that may initially go underappreciated, stakeholders value conversations that catalyze introspection — particularly if it leads to tangible outcomes. Stewarding change presents opportunities for information architects to serve the org’s needs while developing as critical actors in the organization.