Site structure and internal links

Why structure matters for AI answers

Crawlers find pages by following links. A page that nothing links to (an orphan) may never be crawled, however good it is. A page buried five clicks deep behind a vague menu label may be crawled rarely. A page that sits in an obvious place, linked from the navigation and from related pages with descriptive link text, is found quickly and understood in context.

Structure also tells a story about the business. A site whose navigation lists each main service, each location and a clear about page describes the business before a crawler has read a single paragraph. Bay Area AEO's own site is a working example of this: a navigation that lists four industries with their forty niches, and four regions with their cities, so the coverage is visible in the structure itself.

A simple shape that works

Most local business sites need only a few levels.

That is enough for a business with a handful of services. A larger practice can add a level (for example, a dental practice might group pages under general and cosmetic dentistry, with restorative work as a third group), but every additional level should help a customer find something, never just add depth.

Navigation labels

Navigation labels are link text that appears on every page, so they carry weight. Use the words customers use: “Heat pumps”, “Emergency plumbing”, “Private dining”, “Insurance we accept”. Avoid clever or vague labels like “Solutions”, “Experience” or “Discover”. A crawler reading the navigation should be able to list the business's services from the labels alone.

Internal links within the text

Beyond the navigation, link related pages to each other in the body text, with link text that names the destination. A water heater page can link to “tankless water heater installation” and to “our service area in the East Bay”. A bio can link to the services that practitioner provides, and each of those service pages can link back to the bio. These links tell a crawler that the pages belong together, and they let an assistant that lands on one page find the related facts.

Avoid “click here” and “learn more” as link text. They say nothing about the destination.

Service pages as hubs

Each main service page works best as a hub for everything about that service: its FAQs, its price information, the practitioners or crews who provide it, the projects completed, the locations where it is offered. Link out to each, and link back from each. When an assistant retrieves the service page for one narrow question, the links put the answers to the follow-up questions within one step.

Breadcrumbs

Breadcrumbs (a line such as “Home > Services > Water heaters > Tankless” near the top of a page) show where a page sits. They help readers move up a level, and with BreadcrumbList markup they tell machines the same thing. Most platforms can add them automatically.

Readable addresses

Page addresses should be short and descriptive: /services/tankless-water-heaters/ rather than /page?id=4471. A readable address is a small extra statement of what the page is about. Once set, keep addresses stable. When a page must move, redirect the old address to the new one permanently, so links from listings, directories and articles still arrive.

The XML sitemap

An XML sitemap lists every page the business wants crawled. Most platforms generate one automatically at /sitemap.xml. List it in robots.txt and submit it to Google Search Console and Bing Webmaster Tools. Check it occasionally: sitemaps sometimes include pages that should not be there (test pages, old promotions) or miss new ones. This handbook's own sitemap lists all 59 of its pages, and each one is linked from the contents.

The footer as a fact block

The footer appears on every page, which makes it a good place for the facts that should be everywhere: business name, address, phone, hours summary, license number, and links to the contact, about and service-area pages. Keep it in text, consistent with the master record, and resist filling it with dozens of keyword links to city pages.

Structure for several locations

A business with several Bay Area locations should give each its own page, reachable from a locations overview and from the footer. Each location page links to the services offered there, and each service page lists the locations offering it. That two-way linking lets an assistant connect “implants” with “Fremont office” and answer a question that names both.

Finding orphans and dead ends

Over time, sites collect pages nobody links to: last year's campaign page, a service added by a previous manager, a bio for someone who left. Some are worth keeping and linking; others should be removed or redirected. Site-crawling tools list every page and the links to each. A simpler check is to compare the sitemap with the navigation and ask, for each page in the sitemap, how a visitor would reach it.

Dead ends are the opposite problem: pages that link nowhere. A service page with no link to contact, to related services or to its FAQs leaves both readers and crawlers stuck. Every page should offer at least a next step.

The contact page

The contact page is one of the most visited pages on any local business site, and one of the most read by machines looking for hard facts. It should hold the full address with the neighborhood named, directions from the nearest freeway exit or transit stop, parking notes, the phone number as tappable text, the email address, the hours (including how phones are handled out of hours), a booking link if there is one, and a map embed beside the written directions. It should match the listings exactly, because an assistant comparing them will notice any difference.

News, blogs and updates

Many small business sites have a blog that was active for a few months and then stopped. Old posts do little harm by themselves. The risk is the outdated facts they carry: a 2021 post announcing hours that no longer apply, a promotion long ended, a staff member who has left. Review old posts once a year. Update or remove the ones with stale facts, and keep the ones that answer real questions, linking them from the relevant service pages so they are part of the structure. New articles are most useful when they answer a question customers ask, explain a local rule, or describe a real project in detail.

Pages for questions that cross services

Some customer questions cut across the service structure: “How do I prepare my home for a heat pump and solar?”, “What should I do in the first week after a car accident?”, “What does a full kitchen remodel involve from design to inspection?”. A guide page that answers one of these, linking to each relevant service page, is useful to readers and gives an assistant a single source for a multi-part question. Place guides in a clearly labeled section and link each one from the services it covers.

Watching coverage reports

Google Search Console and Bing Webmaster Tools both report which pages they have indexed and which they have skipped, with a reason. Glance at these reports monthly. A new service page that stays “discovered but not indexed” for weeks usually needs more internal links pointing to it, or a clearer place in the navigation.

Linking out

Linking to credible outside sources helps readers and gives context: the licensing board's lookup for the business's license number, the water district's rebate page on a landscaping page, the city's permit page on an ADU page. These links show that the business's claims can be checked, and they place its pages among the sources an assistant already trusts.

Related chapters28. Pages machines can read · 17. Service pages an assistant can quote · 25. LocalBusiness markup · 21. Neighborhood relevance