Can agents find products beyond the first page of your catalog?

A catalog can contain the right product and still give a reader no route to it. The first response lists a few items. The rest appear only after someone presses "Load more." A client that follows links without operating those controls may stop at that first batch.

Keep the scrolling interface if customers like it. Give each batch a direct address and link to the next one as well. That creates a route you can inspect independently of the scrolling code.

Separate crawling from browser interaction

Google documents that its crawlers generally follow anchor links and do not click buttons to reveal more results. Its pagination guidance recommends unique page URLs and sequential links. Those statements describe Google Search, not every agent. A browser agent may operate a button, while another client may only fetch documents. Google pagination guidance.

For your own site, write down the clients you intend to support. "Find a product through category links" and "use the search form in a browser" are separate tasks. A successful search does not tell you whether a reader can traverse the catalog.

Give the next batch an address

Consider this illustrative category route:

<section aria-label="Products">
  <a href="/products/linen-shirt">Linen shirt</a>
  <a href="/products/cotton-shirt">Cotton shirt</a>
</section>
<nav aria-label="Catalog pages">
  <a href="/shirts?page=2">Next page of shirts</a>
</nav>

Requesting /shirts?page=2 should produce the second batch without requiring a visit to page one. Keep the category filter and sorting choice in the URL when they affect which products appear. Opening a shared link in a clean session should preserve the intended collection.

A link needs a real destination. Google's link guidance explains why an anchor with an href is more reliably discoverable by its crawler than a script event on another element. Google's crawlable links documentation.

You can intercept the link in the browser to append results, provided its direct destination remains useful. That is an implementation choice, not a requirement to remove JavaScript. Let the server and browser use the same catalog selection logic so a direct request and a click do not quietly return different batches.

Audit one item you know is missing from page one

Pick a product near the end of a category and record its current URL. Start at the category page in a clean browser, then repeat with a client that only follows the links in each response. Keep the request sequence as evidence.

Check these specific outcomes:

1. Each next-page URL returns the expected batch directly.
2. Product links point to current detail pages.
3. The last batch has a clear end rather than a next link that repeats forever.
4. Refreshing a later page preserves the batch and filters.
5. An invalid page value produces a deliberate response rather than an unrelated collection.

Compare the discovered product identifiers with a small catalog export. This comparison is your own acceptance test; it is not a published benchmark. Choose a stable sample, since adding inventory while the test runs can shift page boundaries.

Keep filtered collections manageable

A category with every possible combination of size, color and sorting can expose many URLs for much the same inventory. Decide which combinations deserve public navigation. Keep useful customer filters, but avoid generating a giant link directory solely for crawlers.

For Google's indexing, each page in a paginated sequence should have its own canonical URL rather than naming page one as canonical. Google also advises against using URL fragments as page numbers. Apply those recommendations within Google's documented scope; neither establishes how every assistant indexes a collection. Pagination URL guidance.

Record what the fix proves

A passing traversal test proves that the selected client found the sampled item through the recorded links. It does not prove an assistant will recommend the product, understand every variant or buy it. Those need their own tests.

Dedica's free readiness checker reports a score and evidence across Discovery, Access, Usability and Payments for a public website. Use it to examine public-site findings alongside your catalog traversal record. A site scan is not evidence that every product in a large inventory was visited.

When reporting the change, include the category URL, sampled product, client and date. "The link traversal now reaches this item" gives a reviewer something they can reproduce.

Sources

• Pagination and incremental page loading, Google Search Central.
• Make links crawlable, Google Search Central.