Thread by @eaton

There are three basic ways to approach content modeling, and I'd argue that each of them have _strengths_ but no single one is philosophically correct or incorrect. Each is just a place to start.

The approach you hear about when a content architect describes ~the process~ is what I think of as "platonic content." You start with a communication or publishing concept and iron out your content types, their properties, etc. in a way that best describes their essential nature.

"We want to do interviews. What _is_ an interview, conceptually? Well, we'll need a subject, perhaps multiple subjects…" and so on.

The second approach, where you'll find many working content strategists, usually starts with a large body of existing "organic" content in a variety of forms. It works towards a model that will allow those different assets to be created, managed, and used consistently.

"We have four CMSs and two newsletters and a nontrivial number of print assets and also a video library… how do we unify them?" The model can't escape contending with the messiness of varied, real-world content, because that's where it starts.

The third angle of approach — often derided, but worth examining — is "design first." You start with visual artifacts and you look at them and say, "OK. What content elements and properties and relationships does this imply? What would need to exist to make this design real?"

At its worst this makes your content model a kind of madlibs exercise. Each design variation becomes a content type, with "holes" cut into the design to be filled by content properties.

That's how the first generation of templated, database driven web sites came to be, but on its own this approach is tremendously brittle — if your design changes, odds are there will be different "holes" to fill and your old data is unlikely to fit them.

Now, I mentioned earlier that none of these approaches is inherently right — a "morally correct" way to model content.

Platonic content modeling can help you build a system of meaning — a formal understanding of what you're making and why and what it _is_ deep down, regardless of where you publish it or who's making it.

…But platonic modeling can also be idealistic — ignoring the messiness of content that doesn't fit its formalisms, leaving out important information the original architects didn't anticipate a need for, or breaking content into needlessly granular components for "accuracy."

The "organic" modeling approach is a pragmatic one, and refuses to ignore the messiness of existing content — it often reveals connections, variations, and important nuance that the platonic approach doesn't anticipate.

…But it's also often blind to the deeper patterns beyond the existing data. At its worst, it "lifts and shifts" existing structures, ignoring opportunities for consolidation and improvement by only considering what the content IS rather than what it should or could be.

And the design-first approach, well. Loads of us in the content strategy, IA, and dev worlds are familiar with the downsides of that. We rail against it, sometimes even insisting that content architecture and modeling should avoid the "corrupting invluence" of presentation

But good design isn't just decorative icing spread on top of the content cake: it anchors the content in context, and often reveals information that NEEDS to be captured in the content model to make effective presentation possible.

When design is ignored, "platonic" content models often accumulate weird, editorially controlled buffer layers — secondary wrapper content types (or, God help us, secondary CMSs) that act as sin eaters for all the metadata and relational decisions that drive the UX.

Do editors want some stories on the front page to stand out from the rest? If the content model doesn't have some way to capture that intent, and simple rules ("always the third story from the top") are insufficient, well… you're SOL.

The idea here isn't that we're screwed no matter which path is taken.

(Okay, maybe it is, but that's just modernity.)

The point is that the three approaches compliment and inform each other. Different disciplines favor different approaches, but each is incomplete in isolation.

I usually advise folks to start with the angle with the most certainty — whether that's an existing inventory, stakeholder buyin on core messages and mediums, a set of approved designs, etc. Then cycle through the other perspectives, using them to identify gaps or conflicts.

That usually requires bringing in other people — the mobile team to talk about how they want the iOS app to present the content, SEO experts to talk about underlying metadata needs, marketing and comms to discuss the "whys" behind the goofy variations in existing material, etc.

That iterative, cross-disciplinary approach isn't just a CYA move to ensure stakeholders are heard; it's the only way to really understand what the content is, and what it needs to be.