Cache plugin reviews read like feature lists because nobody runs them long enough to see what breaks. We installed WP Rocket on a live WooCommerce store (600 products, wholesale role pricing, EU traffic) and kept it there for 90 days. This is what actually happened — including the two defaults that broke pages and the settings we changed within the first week.
The before/after numbers
Measured with the same protocol as our cache plugin comparison, on production traffic patterns:
| Metric | Before | After (week 1) | After (90 days) |
|---|---|---|---|
| Cached LCP (home) | 1.94s | 0.71s | 0.68s |
| TTFB (cached, p75) | 610ms | 96ms | 91ms |
| Uncached LCP | 1.94s | 1.88s | 1.71s |
| Server CPU load | baseline | -18% | -22% |
Cached visits improved exactly as advertised. Note what did not move much: uncached LCP — because caching cannot fix a heavy theme, as we keep repeating in the theme diagnostic guide. The small uncached gain came from WP Rocket’s font preloading and DNS prefetch, not caching.
The defaults that broke things
- Minify/combine CSS — broke the checkout’s styled shipping-methods block on day 2. Excluding those files fixed it; the exclusion UI is genuinely good. Lesson stands from our form plugin testing: third-party plugin assets need exclusions, not global rules.
- Lazy-load images on category pages — pushed LCP up on archive pages because the first visible product images got delayed. We disabled lazy-loading above the fold (WP Rocket now auto-handles this on newer versions, but verify your setup).
Settings we changed and why
- Preload turned on — the real TTFB win at scale: cache warms as pages are visited, p75 stays near best-case.
- Cache lifespan extended to 14h — the store barely changes; default 10h caused occasional rebuild spikes.
- Excluded the wholesale price page — role-varied content is the scenario analyzed in the role-content test; one exclusion beat subtle caching bugs.
What WP Rocket does not do
It is not a CDN (bring your own — we use Cloudflare, configured by the user), not an image optimizer, and not a theme fix. Sites expecting one plugin to fix everything stay disappointed. Its job — page cache, preload, sensible minification — it does reliably.
Verdict after 90 days
Worth the ~$59/yr on stores with mixed content and stakeholder patience constraints. For a simple site, free Cache Enabler measures within 0.03s. Either way, measure before/after with real pages — the free audit gives you the baseline.
After 90 days on a live store: yes, for mixed-content sites needing reliable exclusions and preload. Cached TTFB dropped from 610ms to ~90ms. Simple sites can match most of it with free plugins. CSS combination broke checkout styling and default lazy-loading delayed above-the-fold images. Both fixed via exclusions/toggles within a week. Test checkout flows after enabling minification. Marginally — font preloading and DNS prefetch helped ~0.2s. Cache plugins cannot fix heavy themes; the uncached tax is a theme problem (see our diagnostic guide). Yes — preload keeps the cache warm so real visitors rarely hit uncached pages. It was the biggest sustained win in our 90-day run.Is WP Rocket worth it for WooCommerce?
What broke when we installed WP Rocket?
Does WP Rocket improve uncached page speed?
Should preload be enabled in WP Rocket?
