WordPress sites before and a year after adding a caching plugin, phones
| Measure | Added a caching plugin | No caching plugin either time |
|---|---|---|
| Cases | 195,378 | 2,504,716 |
| Good LCP before | 41.5% | 58.5% |
| Good LCP after | 57.8% | 59.4% |
| Slow sites that reached good LCP | 39.8% | 19.7% |
| Pass CWV before | 36.3% | 52.6% |
| Pass CWV after | 51.7% | 54.2% |
| Median LCP change | -400 ms | 0 ms |
Source: PageSpeed Matters, Nov 2026. Site-years with phone field data at both ends.
Summary
"Install a caching plugin" is the first answer to almost every slow WordPress site question. We wanted to know how often it works for real visitors: when a slow site adds one, how often does its loading on phones become good within a year?
About 4 times in 10. Of slow WordPress sites that added a caching plugin, 39.8% reached good loading (LCP) within a year, twice the 19.7% of slow sites that never had one. Across all 195,378 cases, the share passing Core Web Vitals rose from 36.3% to 51.7%. The other 6 in 10 slow sites needed more than a plugin.
What we measured
We used HTTP Archive mobile crawls of WordPress homepages from October 2024, October 2025 and October 2026, and found sites where a caching plugin (such as LiteSpeed Cache, WP Rocket or W3 Total Cache) appeared on the homepage in one crawl but not in the one a year before. We matched each site to Chrome UX Report phone field data at both ends. That gave 195,378 cases. The comparison group is 2,504,716 WordPress site-years with no caching plugin at either end.
A slow site is one without good LCP before (LCP over 2.5 seconds at the 75th percentile of phone visits). A site passes Core Web Vitals when LCP, INP and CLS are all good at the 75th percentile.
Limits: this is before and after, not a controlled test. Sites that add a caching plugin often change hosts, themes or images at the same time. Caching that runs on the server without a plugin is not counted.
Twice as many slow sites reached good LCP
Of the slow sites that added a caching plugin, 39.8% had good LCP a year later. Of slow sites that never had one, 19.7% got there, so adding a plugin went with twice the rate.
The median site that added a plugin got 400 ms faster at loading its main content. The median site without one did not change.
The gain holds without LiteSpeed Cache, which hosts often switch on. It makes up 78,944 of the 195,378 cases. Leaving those out, 37.6% of slow sites still reached good LCP.
39.8%
Share of slow WordPress sites that reached good LCP on phones within a year of adding a caching plugin, against 19.7% without one.
Pass rates rose 15.4 points
Across all cases, the share of sites passing Core Web Vitals rose from 36.3% before the plugin to 51.7% a year after, a gain of 15.4 points. Good LCP rose from 41.5% to 57.8%.
Sites without a caching plugin gained 1.6 points over the same years, from 52.6% to 54.2%.
Note where the adders started: 16.3 points behind the sites that never had a plugin on pass rate. Owners likely reach for caching when a site is already slow, and a year later the adders sit 2.5 points behind the sites that never had a plugin.
Source: PageSpeed Matters. HTTP Archive and CrUX phone data, Oct 2024 to Sept 2026.
Adders started slower, which leaves more room to gain.
6 in 10 slow sites needed more than a plugin
The flip side of 39.8% is that most slow sites that added caching still did not have good LCP a year later. Caching makes the server answer faster, but it does not shrink a heavy hero image, remove render-blocking scripts or fix a slow database query on uncached pages.
Which plugin was added made only a few points of difference: WP Rocket cases reached good LCP 45.0% of the time and LiteSpeed Cache cases 42.7%. The setup, and the work around the plugin, likely matter more than the brand.
What this means for your site
If your site is slow and has no caching, add it: in our data it went with double the rate of reaching good loading within a year. Then check whether it was enough.
- Turn on page caching first. Serving saved pages is the part that cuts server time.
- Check the main image. Caching does not make a large hero image smaller or load it sooner.
- Defer what can wait. Scripts that block the first paint still block it on a cached page.
- Measure field data after a month. Lab scores move at once; phone field data shows what visitors got.
When caching alone is not enough, our WordPress speed optimization service tunes the cache rules, shrinks and prioritizes the main image, defers scripts that block the first paint and fixes slow uncached pages, then tracks phone field data until the site passes. After the site is fast, our WordPress support services keep the caching plugin updated, clear and test the cache after every theme or plugin change and watch for pages that drop out of it, so the gain does not fade over the next year.
Sources
- HTTP Archive mobile crawls of WordPress homepages, October 2024, October 2025 and October 2026, with Chrome UX Report phone field data (October 2024, October 2025, September 2026), analyzed November 2026
- web.dev, Largest Contentful Paint: https://web.dev/articles/lcp
Frequently asked questions
Do caching plugins actually speed up WordPress?
In real-visitor data, yes for many sites: 39.8% of slow WordPress sites that added a caching plugin reached good LCP within a year, against 19.7% without one. This is before and after across 195,378 cases, not a controlled test.
Is WP Rocket better than LiteSpeed Cache?
Slightly, on this measure: 45.0% of slow sites that added WP Rocket reached good LCP within a year, against 42.7% for LiteSpeed Cache. LiteSpeed is often switched on by the host, so the groups differ.
Why is my WordPress site still slow after installing a caching plugin?
Most slow sites were: about 6 in 10 still lacked good LCP a year after adding caching. Caching speeds up the server's answer but does not fix heavy images, blocking scripts or slow uncached pages.