Internal Linking for Directories: The Structure Google Rewards
Directories have thousands of similar pages and usually link them badly. The hub structure that works, how to link entries to each other meaningfully, and the patterns to avoid.
Internal linking is the cheapest SEO work available and directories get it wrong more consistently than any other site type, because the linking is generated rather than written.
The symptom is familiar: hundreds of entry pages that Google crawls once, indexes reluctantly, and then largely ignores.
Why directories struggle with this
A blog links naturally. Writers reference earlier articles because the argument requires it, and the anchor text is whatever made sense in the sentence.
A directory links structurally. Every link is produced by a template, which means every link is either right across ten thousand pages or wrong across ten thousand pages. There is no middle.
Get the templates right once and you have solved it permanently. Get them wrong and no amount of manual work compensates.
The structure that works
Three tiers, each with a job.
Homepage. Links to top categories only. Not to entries, not to every facet. Its job is to pass authority into the category layer and establish what the directory covers.
Category pages. Your primary SEO surface. Each links to its entries, to sibling categories, and back up to the homepage. This is where most of your internal link equity should be concentrated, because these are the pages ranking for your main clusters.
Entry pages. Link back up to their category, sideways to genuinely related entries, and out to the business. This is the tier that usually fails, and the sideways links are where the failure sits.
Collections sit alongside categories as a second hub layer. A collection grouping every entry with a shared trait, gunsmiths offering partner-armoury delivery for instance, is both a good landing page for comparative queries and a useful link hub.
Related entries, done properly
Most directories generate related links by category. Every entry links to five other entries in the same category, usually the most recent ones or a random selection.
That is nearly worthless. The links carry no information, the anchor text is just a business name, and a visitor gains nothing from following one.
What works instead is relating on your typed attributes. If your schema captures the properties a buyer decides on, relate on those.
For gunsmiths: other gunsmiths handling the same regulatory weapon category, or other entries within the coverage radius that also offer partner-armoury delivery. For nurseries: others with the same age range and comparable hours.
The difference is that the link now answers a question the visitor already has, which is what makes it worth following and what makes it meaningful as a signal.
This is another case where the data model does the work. Without typed attributes, related-entry linking has nothing to relate on.
Anchor text
Directories default to business names as anchor text, which tells search engines almost nothing.
Where the template allows it, add context. "Sedation dentistry in Lyon" carries more information than the practice name alone. It also reads better, which matters because a link nobody clicks is a weaker link.
Do not over-optimise. Repeating the same exact-match phrase across four hundred internal links is a pattern, and not a good one. Vary naturally.
Crawl depth
Every page should be reachable within three clicks from the homepage. Four at the outside.
Directories break this quietly through pagination. If a category has 200 entries at 20 per page, the entries on page eight are seven clicks deep and effectively invisible to crawlers.
Fixes, in order of preference:
Add facet hub pages that break large categories into meaningful subsets, each linking directly to its entries. This solves crawl depth and creates useful landing pages at the same time. Only create them where five to ten real entries sit behind them.
Increase entries per page where performance allows.
Link to the full range in the sitemap, which helps discovery but does not fix the internal link equity problem.
Pagination itself should be crawlable, with real links rather than JavaScript-only controls.
What to avoid
Linking every page to every page. Some directories put a full category list in the footer with every subcategory expanded. That dilutes every link on the site and creates a flat structure with no hierarchy.
Nofollowing internal links to control equity flow. This has not worked as intended for many years and mostly just wastes the link.
Orphan pages. Entries reachable only through search are not reachable at all as far as crawling is concerned. Every entry needs a path from a category or collection.
Linking to thin pages. If a facet page has two entries behind it, linking to it from four hundred entry pages tells search engines you consider it important. You do not.
Breadcrumbs that skip levels. They should mirror the actual hierarchy, and they should be marked up with BreadcrumbList.
The blog connection
If you publish content alongside the directory, it should link into the set rather than sitting beside it.
An article about choosing a gunsmith should link to the relevant category and to two or three specific entries that illustrate the point. That passes authority from content that attracts links into pages that convert.
The reverse also matters. Category pages should link out to the relevant guide, which gives an uncertain visitor somewhere to go instead of leaving.
Most directories run their blog as an island. It is the easiest link equity to redirect and almost nobody does it.
A practical audit
Half an hour, once, then quarterly.
- Crawl the site and check maximum click depth. Anything past four needs a hub.
- Find orphan pages with no internal links pointing at them.
- Check where equity concentrates. If your footer has 200 links, that is where it is going, and it should not be.
- Sample ten entry pages and read the related links. Would a visitor follow any of them? If not, the relation logic is wrong.
- Check whether your blog links into the directory and whether the directory links back.
Each of these is a template fix rather than a page fix, which is why the effort is small and the effect is site-wide.
FAQ
How many internal links per page?
No fixed number. Enough to be useful, few enough that each one means something. A footer with 200 links fails both tests.
Should entry pages link to each other?
Yes, when the relation is meaningful. Random same-category links are close to worthless.
Do internal links still matter with AI search?
More, if anything. Structure and hierarchy help systems understand what your site is authoritative about, which is what citation depends on.
Should I use a sitemap instead?
Sitemaps aid discovery. They do not pass authority. You need both.
What about links out to the listed businesses?
Keep them. Linking to relevant external sites is normal and expected. Consider rel attributes appropriate to your policy for unverified or user-submitted entries.
Describe your niche and see the schema an agent designs →