Shipping Entity Hub: From Concept to Enterprise Product

Entity Hub started as a feature request and became one of Schema App's core products. Enterprise customers were managing thousands of entities across their content but had no centralized way to see what entities existed on their site, how they were connected, or whether they were performing well in AI search. They were flying blind.

What the product needed to do

We scoped Entity Hub into three pieces: Entity Manager for editing and organizing entities, Entity Reports for a site-wide view of entity coverage and gaps, and Entity Performance Analytics for tying Google Search Console data to individual entities. The first two were relatively straightforward engineering. The third was where things got interesting.

Google Search Console gives you data at the URL and query level. It doesn't know anything about entities. To show clicks, impressions, and CTR at the entity level, we had to build a mapping layer that connects GSC URL data to the entities present on each page, weighted by entity prominence. That mapping had to run at scale (some customers have 10,000+ pages) and update regularly as both the content and the search data change.

The hard part was the data model

Entities aren't pages. A single entity can appear on multiple pages. A single page can contain many entities. When a customer asks “how is my Organization entity performing in search?” the answer depends on every page where that entity appears, weighted by how prominent it is on each page.

We went through three iterations of the attribution model before landing on something customers found useful. The first version was too simple (just count which pages an entity appears on). The second was too complex (a weighted model that nobody could explain to their CMO). The third was a middle ground: clear attribution with a simple prominence score, and the ability to drill down into per-page detail if needed.

What enterprise customers actually wanted

We assumed customers would spend most of their time in Performance Analytics. They didn't. The most-used feature turned out to be Entity Reports, specifically the gap analysis view that shows which topics they should have entities for but don't. Enterprise SEO teams use this to make content strategy decisions and justify new content investments to leadership.

The lesson: don't assume the most technically impressive feature is the most valuable one. The GSC integration was hard to build and customers appreciated it, but the simple gap report is what they open every week.

Shipping to enterprise

Entity Hub launched as an add-on to Schema App's existing platform. We did a phased rollout, starting with three design partners who gave us weekly feedback. That feedback loop was critical. Without it we would have shipped a version that was technically correct but didn't match how SEO teams actually work.

The product is now live with enterprise customers and it's become a core part of how they think about AI search readiness. The entities in their Content Knowledge Graph aren't just metadata anymore. They're a managed, measured asset. That shift in how customers think about their data is probably the biggest win.