Diagnosing a Slow WooCommerce Store: Six Measurements Before You Change Anything
Server time, anonymous testing, page types, query counts, page weight and cache behaviour. A measurement sequence that decides which kind of performance project you actually have.
“The site is slow” is not a diagnosis. It is a complaint. Between that sentence and a fix there are five or six measurements, and skipping them is why so many performance projects replace a plugin that was never the problem.
This is the sequence we run before changing anything on a WooCommerce store.
Step 1: separate server time from browser time
The first measurement is not a page score. It is: how long does the server take to produce HTML, with no images, scripts or CSS involved?
- Under 200ms — the server is fine; the problem is on the page
- 200–800ms — borderline; caching and queries deserve a look
- Over 800ms — nothing you do in the browser will fix this; the server is the problem
Test with a plain request from a location near your customers, not from your office. A store on a European host tested from China will look catastrophically slow and be perfectly fine for its actual users.
This one number determines which half of this article applies to you.
Step 2: test as an anonymous visitor
Almost every “it’s fast for me” report comes from testing while logged in as an administrator, with the admin bar, uncached pages, and often a development plugin active.
Use a private window, logged out. If the difference between logged-in and anonymous is large, caching is working and the slow experience is admin-only — a different, smaller problem.
If both are slow, caching is either not working or not covering the pages that matter.
Step 3: find out which page types are slow
WooCommerce stores have four page types with completely different performance profiles:
| Page type | Typical cause of slowness |
|---|---|
| Home | Too many product queries, unoptimised sliders, large hero images |
| Category / shop archive | Faceted filters running live queries, sorting on unindexed meta, deep pagination |
| Product | Variation queries, related products, review aggregation, gallery images |
| Cart / checkout | Uncacheable by design, so every query is exposed; shipping and tax calculations per request |
Measure each type separately. A store where the product page takes 3 seconds and checkout takes 8 has a different problem than one where everything is uniformly slow, and the fixes barely overlap.
Category pages and checkout are where most WooCommerce stores actually lose money, and both are frequently not measured at all because the homepage looks fine.
Step 4: count the queries
A normal WooCommerce product page runs 40 to 80 database queries. Above 150, something is looping — usually a plugin loading its data per product, or a related-products block querying in a loop.
Enable query logging on staging, load one category page and one product page, and look for:
- The same query repeated dozens of times — the classic N+1 pattern
- Queries on
postmetafiltering by value without an index — these get slow fast as the catalogue grows - Queries taking over 100ms each, regardless of how many there are
The total query count matters less than duplicates. Ten distinct queries is fine; one query run 200 times is the bug.
Step 5: weigh what the browser downloads
Now, and only now, look at the page itself. Three numbers:
- Total transfer size — over 3MB on a product page usually means unoptimised images
- Number of requests — over 100 often means a page builder and several plugins each loading assets everywhere
- Largest single resource — usually the hero image or a render-blocking stylesheet
Images are the most common culprit and the easiest to fix. A 2400-pixel JPEG served at 600 pixels is a saving waiting to happen — the same problem described in the image migration notes.
Step 6: check what caching is actually doing
Print the response headers for an anonymous request to a category page. You are looking for evidence that a cache served it — a hit header from your page cache, or a CDN cache status.
Three common states:
- Cache miss on repeat requests — the cache is misconfigured or the page is excluded
- Cache hit, but slow — the problem is upstream: DNS, TLS, or the CDN edge being far from the customer
- No cache headers at all — no page cache is active, or it is bypassed for all visitors
Cart and checkout are correctly uncacheable. If they are being cached, that is a security problem, not a performance one — stop and fix it first.
Step 7: decide before you act
After these six measurements you will usually find that one of these is true:
The server is slow. Fix hosting, PHP version, database, or object cache. Not plugins.
The pages are heavy. Fix images, fonts, scripts. Usually a week of work with a large win.
Specific pages are pathological. Usually filters, variations or a plugin. Fix those pages, leave the rest alone.
Nothing is wrong technically, but perceived speed is poor. Layout shift, slow font loading, a spinner that lingers. Fix perceived performance, not throughput.
Each of these leads to a completely different project. Choosing between them by measurement takes a few hours; choosing by intuition takes months and often fixes the wrong thing.
If you want this run as a diagnostic rather than a rebuild, that is how we approach speed optimisation — the first deliverable is a measurement report, not a list of plugins to install.