WooCommerce development

Where WooCommerce stores break#

WooCommerce works with almost any configuration up to about a thousand products. After that, three things start to break.

Product filtering. Attribute-based filters become queries against wp_postmeta. That table gains dozens of rows per product, and without a suitable index a two-condition filter can take seconds.

Search. WordPress's default search runs LIKE %...%, which cannot use an index at all.

Uncacheable pages. Cart, checkout and account are never cached. If your site is only fast because of page caching, those three pages — exactly where revenue happens — are still slow.

Migration is rarely the answer#

The common advice at this point is "move the store to X". That is almost always more expensive and riskier than fixing the problem where it is.

-- A composite index matching the store's real query pattern, rather than
-- relying on the default meta_key index, which was not built for this.
ALTER TABLE wp_postmeta
  ADD INDEX meta_key_value_post (meta_key(32), meta_value(32), post_id);

For very large catalogues the better answer is usually a flat lookup table holding only the filterable fields, kept in sync through product save hooks. The filter query becomes a single SELECT instead of several joins against postmeta.

Checkout is where money leaks#

Every extra field in the checkout form costs a percentage of orders. What I typically do:

• Remove fields that are not genuinely needed.
• Validate postcode and phone on the client so the customer does not bounce back with an error after submitting.
• Strip unnecessary scripts from the checkout page — it needs neither a slider nor an Instagram widget.