It's Not Magic. It's Metadata: Aligning Content Strategy to Business Outcomes
I presented "It's Not Magic. It's Metadata! Aligning Content Strategy to Business Outcomes" at Adobe DITA World on June 25, 2026. The talk was for content leaders who already have a content strategy, a CMS, a data model, and an experienced team, but still can't answer the questions upper management is asking: publish faster, prove AI readiness, and justify continued investment. A metadata strategy is what translates a content strategy's plan into execution, using governance to catch drift before it compounds, automation to remove manual friction, and measurement to make content's business impact visible to leadership.
I framed the gap in terms of three failure modes: drift, friction, and invisibility. Drift is what happens as content grows and metadata standards, staff, and systems change without anyone tracking the divergence; governance closes that gap by removing ambiguity about which values apply, where, and by whom. Friction is anything still done by hand; automation is how teams keep up with content demand without depending on individual heroics. Invisibility is the failure to connect content metrics to the business outcomes leadership actually tracks, which is why investment dries up even when the work is good. To make the gap concrete, I walked through a composite enterprise example: 150,000 topics, an experienced content team, eight-language delivery, and a functioning taxonomy with room to grow into ontologies and knowledge graphs, that still can't confirm its content is ready for LLM consumption because there's no decision framework for which metadata values apply where, no way to test how metadata performs, and no path from content metrics back to business outcomes.
The session closed on why generative AI raises the stakes: AI is an amplifier, so messy content without a metadata layer just produces more of it, faster, with no better chance of matching what a user actually asked. I covered how DITA's own semantic markup already functions as metadata for this purpose, using a task topic's prerequisite element as the example: it tells an LLM exactly where to look for what has to happen before a step, instead of requiring it to search the whole topic. The talk ended with a direct challenge to the audience: name one place where the gap between your content strategy and your metadata is costing you, and figure out how to measure that cost before deciding how to fix it.
Curious where your organization's metadata strategy has a gap? Set up a conversation with Amber.
If you're still building the controlled vocabulary this talk assumes you already have, this webinar on taxonomy definition and governance walks through it from the ground up.
Not sure how much metadata is already hiding in your content? This practical checklist for finding and applying it is a good next step.
For the deeper dive on ontologies and knowledge graphs this talk only touched on, see this roundtable on what makes content genuinely intelligent.
