In many publishing systems an article is essentially a page: a headline, a body of formatted text and some images placed where they look right on the website. That works for the website. It works less well everywhere else, because every other channel has to unpick the page to find the parts it needs.
Structured content takes the opposite approach. The article is stored as distinct, labelled parts, and each channel assembles the parts it needs.
What structure looks like in practice
- Headline, short headline and social headline, which are rarely the same
- Standfirst and summary, for listings, newsletters and search
- Body text, with quotes, galleries and videos stored as items in their own right
- People, places, organisations and topics, held as references rather than free text
- Image, caption, credit and rights, kept together and reusable
Design the model with the people who write
A content model designed only by developers tends to be thorough and hard to use. Editors face long forms of optional fields, and the structure decays as people skip what they do not understand. The model works when the people who write and edit agree what each field is for, and the editing screen asks only for what is needed.
Structure is only valuable if editors fill it in, so design it for the people who do.
Headless, or not
Structured content is often discussed alongside headless CMS platforms, where the content store and the website are separate. Headless suits publishers with several channels, or a front end that needs to change quickly. But structure does not require headless: Drupal and WordPress can both hold well-structured content and still render the website themselves. The decision should follow the channels you actually publish to.
Plan the move from the old model
- Audit how existing body text is really used, including embeds and shortcodes
- Decide which parts can be extracted by script and which need an editor's eye
- Give the most-read and most-linked articles the most care
- Keep every old URL working, so links and search rankings carry over
The content model is the part of a project that is hardest to change later. We design it in discovery, test it against real articles from your archive, and only then build on it.
