Directory sites

A directory is a database with a website on top. Most problems come from the data model, so that is where we start.

Data first

We model the entities and their relationships before designing a page: listings, locations, categories and the attributes people filter on. A school directory, for instance, needs curricula, age ranges, fees by year and campuses, and a school can have several campuses in different districts.

Page types

• Listing pages, one per entity, with the facts a reader needs.
• Location pages for each country and city, with a list of listings and a short guide.
• Category or filter pages for the combinations people search for.
• Editorial guides that answer questions the listings cannot.

Thin pages

Directories can generate thousands of pages with almost nothing on them. Search engines treat those poorly and readers bounce. We set minimum content rules, such as not publishing a city page until it has enough listings, and noindex filter combinations that add no value.

Structured data

We mark up pages with schema.org types that fit what is on the page, such as BreadcrumbList, CollectionPage and ItemList for location pages. Our notes on structured data for a university directory go into detail.

Keeping it current

Data goes stale. We add a "last checked" date to each listing and build simple tools for updating fees and contact details in bulk.