WordPress site-years that changed vs kept their host: phone field data at the start and end of the year
| Measure | Changed host | Same host |
|---|---|---|
| Site-years | 39,432 | 1,140,028 |
| Good LCP at the start | 55.3% | 60.2% |
| Good LCP a year later | 64.2% | 61.3% |
| Slow sites that gained good LCP | 42.2% | 22.2% |
| Fast sites that lost good LCP | 18.1% | 13.0% |
| Server response (TTFB) improved | 52.7% | 38.1% |
Source: PageSpeed Matters, Nov 2026. HTTP Archive WordPress homepage crawls with CrUX phone data, October 2024 to September 2026; TTFB over sites with data at both ends.
Summary
Changing host is a common step for WordPress owners whose site feels slow. We checked what real visitors on phones saw in the year after a move, for sites that started slow and sites that started fast.
For slow sites, a move often paid off. 42.2% of slow WordPress sites that changed host had good LCP a year later, nearly twice the 22.2% of slow sites that stayed put. But moves cut both ways: 18.1% of fast sites that changed host lost good LCP, against 13.0% of fast sites that stayed. A host move is a real chance to get faster, and a real chance to get slower if it is not checked.
What we measured
We used HTTP Archive crawls of WordPress homepages in October 2024, October 2025 and October 2026, with Chrome UX Report phone field data for October 2024, October 2025 and September 2026. Each site-year runs from one October to the next. A site changed host when the hosting provider detected on it at the start differed from the one at the end. We can detect 18 named hosts.
Good LCP means the largest element loads in 2.5 seconds or less at the 75th percentile. Slow sites started the year without good LCP; fast sites started with it. We also checked whether server response time (TTFB) improved, among sites with TTFB data at both ends.
Two limits matter. A host move often comes with other changes, such as a new theme, caching setup or rebuild, and we cannot separate them. And about a fifth of changes are a second detected host appearing or disappearing rather than a clean swap.
42.2% of slow sites that changed host got good LCP
Of the slow site-years that changed host, 42.2% had good LCP a year later. Of the slow sites that stayed on the same host, 22.2% did.
So a slow site that moved was nearly twice as likely to reach good LCP within the year. Across all 39,432 site-years with a host change, the share with good LCP rose from 55.3% to 64.2%, a gain of 8.9 points. The 1,140,028 site-years on the same host rose from 60.2% to 61.3%.
42.2%
Share of slow WordPress sites that had good LCP on phones a year after changing host, vs 22.2% of slow sites that stayed.
18.1% of fast sites that moved lost good LCP
The risk runs the other way for sites that were already fast. 18.1% of fast site-years that changed host had lost good LCP a year later, against 13.0% of fast sites that stayed.
That is more than 1 in 6 fast sites losing good LCP after a move. The likely causes are the usual ones: a new server further from visitors, caching not set up on the new host, or a rebuild that came with the move.
Server response improved for 52.7% of movers
Among the 36,838 host-changing site-years with TTFB data at both ends, 52.7% had a faster server response a year later. Among sites on the same host, 38.1% did.
Server response is where a host change acts most directly, and it moved the right way for just over half of movers. Caching and server tuning on the new host are likely what separate the movers that gain from the rest.
Source: PageSpeed Matters. CrUX phone TTFB, sites with data at both ends.
Counted only where TTFB data exists at both ends.
What this means for your site
If your WordPress site is slow, moving host was often followed by good LCP in our data. If it is already fast, a move needs more care, not less.
- Measure before you move. Record field LCP and TTFB so you can judge the new host on real visitors.
- Match caching and server settings. Page caching, object caching and compression may need setting up again on the new host.
- Pick a data center near your visitors. Distance adds to every request.
- Check again a month after launch. Field data shows within weeks whether the move held your speed.
When you change host, our WordPress migration services record field speed before the move, carry over caching and server settings, switch DNS without downtime and check LCP again after launch, so the move is set up to land among the sites that got faster. On the new server, our WordPress speed optimization service tunes page and object caching, trims slow plugins and fixes the largest image on key templates, so server response gains turn into good LCP for real visitors.
Sources
- PageSpeed Matters analysis of HTTP Archive crawls of WordPress homepages, October 2024 to October 2026, with Chrome UX Report phone field data, analyzed November 2026
- web.dev, Time to First Byte: https://web.dev/articles/ttfb
Frequently asked questions
Does changing WordPress host make a site faster?
For slow sites it often goes with a faster site. 42.2% of slow WordPress sites that changed host had good LCP on phones a year later, vs 22.2% of slow sites that stayed. Host moves often come with other changes, so this is not the host's effect alone.
Can moving host make a WordPress site slower?
Yes. 18.1% of fast WordPress sites that changed host lost good LCP within a year, vs 13.0% of fast sites that stayed. Caching and server settings that are not carried over are a likely cause.
How often does a new host improve server response time?
Among 36,838 host-changing site-years with TTFB data at both ends, 52.7% had faster server response a year later, vs 38.1% of sites on the same host.

