Large-catalogue search for WooCommerce

Where a growing catalogue actually hurts

WooCommerce stores catalogues in two tables: wp_posts holds one row per product, and wp_postmeta holds everything else — price, SKU, stock, and every attribute of every variation. That second table is the multiplier: a modest 3,000-product store with variations can carry hundreds of thousands of postmeta rows.

Writing to those tables scales fine. Scanning them doesn’t — and search and filtering are scans. That’s why the pain of scale arrives in a particular order: admin product lists slow a little, category pages a little more, and search first and worst, because a leading-wildcard LIKE can’t use an index at all. It’s rarely a cliff; it’s a slope that steepens with every import.

Worse, the failure is self-concealing. You rarely search your own store the way shoppers do — you go straight to products from the admin. The people who feel the ten-second search page are the ones who never tell you: they close the tab.

What “large” means for WooCommerce search — honestly

There’s no universal number, because hosting, plugins and variation counts vary wildly. As a rule of thumb from stores we see: native search starts feeling sluggish in the low thousands of products, and by five figures it’s genuinely painful — especially once SKU or attribute plugins add postmeta joins to every query. If your store is on the slope, you already know: search is the page you apologise for.

The external-index answer

The architecture that fixes scale is the one built for it: your catalogue syncs to a dedicated search index, and every query — every keystroke, every filter click — runs there instead of your database. Your server’s job shrinks to what it was sized for: rendering pages and taking orders.

• Search cost stops tracking catalogue size on your host. The engine is built to answer catalogue-scale queries typically in around a tenth of a second; your hosting does none of the search work, at 500 products or 10,000.
• Filtering scales with it. Facet counts — the expensive part — come back with results in one round trip. See faceted search for why that matters more as catalogues grow.
• The first sync is built for big. Your catalogue queues through the plugin’s outbox and every chunk is confirmed by the engine — an interrupted sync resumes rather than restarts, and search goes live as products land in the index, not after the last one.

Sized in the open

Plans are priced by catalogue size and nothing else — flat, so a busy month costs what a quiet one does. Ultimate covers up to 10,000 search results at $19.99/mo. Beyond that is Enterprise: limits sized to your catalogue — including unlimited — agreed with a human and invoiced, not metered. The full ladder, including what happens if you outgrow a plan (search keeps working; new products pause until you upgrade — we never purge an index), is on the pricing page.