Speed & performance · Guide
How to Tell If Your Host Is Slowing WordPress Down
Short answer
Your host is likely slowing WordPress down when time to first byte stays over 1.8 seconds on several pages, even after caching is working and heavy plugins are ruled out. Check PageSpeed Insights field data, the Waiting time in Chrome DevTools, and Tools, Site Health, Info in WordPress. If the server is slow on a plain page too, the host or plan is the main suspect.
Key takeaways
- A TTFB of 0.8 seconds or less is good, and over 1.8 seconds is poor, according to web.dev guidance checked in November 2026.
- PageSpeed Insights field data shows what real visitors experienced over the last 28 days, not one lab run.
- WordPress.org recommends PHP 8.3 or greater, MariaDB 10.11 or greater, or MySQL 8.4 or greater, as of November 2026.
- A cached page that is still slow, or a slow plain HTML file, points at the server rather than WordPress content.
- PHP 8.2 stops receiving security fixes on 31 December 2026, so a host stuck on it needs a plan.
How can you tell if the host is the problem, not the site?
The host is the main suspect when the server is slow before WordPress content even starts to load.
Start with time to first byte, known as TTFB. It measures the wait from the start of navigation until the first byte arrives. If that wait is long on a simple post and on a plain file, the delay sits with the server.
A slow site can still have a fast host. Large images, render-blocking scripts, and a heavy theme slow the page after the first byte arrives. That is why you must split the load into two parts: the server wait, and everything after it.
Use this order for the diagnosis:
- Measure TTFB for real visitors with PageSpeed Insights.
- Repeat the check in Chrome DevTools for one page load.
- Read the server facts in WordPress Site Health.
- Test a cached page and a plain file to separate caching from server speed.
If the server wait stays high through all four steps, you have host evidence, not a guess. If it drops when caching works, the fix may be a cache setting, not a new host.
What is TTFB and what counts as slow?
TTFB is the time from the start of navigation until the first byte of the response arrives.
TTFB includes redirects, DNS lookup, the connection, TLS negotiation, and the server work before it sends the first byte. web.dev describes it as a foundational metric because every later metric waits for it. Nothing on the page can load until the HTML starts to arrive.
web.dev gives a rough guide, checked in November 2026: most sites should aim for a TTFB of 0.8 seconds or less, and a value over 1.8 seconds is poor. The same guide says to judge the 75th percentile of users, so at least 3 in 4 visits meet the mark.
TTFB is not one of the three Core Web Vitals. It matters because it eats into Largest Contentful Paint, known as LCP. Google's LCP guide calls 2.5 seconds or less good and over 4.0 seconds poor, also at the 75th percentile.
| Metric | Good | Poor |
|---|---|---|
| TTFB | 0.8 seconds or less | Over 1.8 seconds |
| LCP | 2.5 seconds or less | Over 4.0 seconds |
| INP | 200 ms or less | Over 500 ms |
INP, or Interaction to Next Paint, measures how fast the page reacts to a tap or key press. A good TTFB with a poor INP points at scripts and main-thread work, not at the host. Keep that split in mind before you blame the server for every slow score.
How do you measure TTFB with PageSpeed Insights?
Open PageSpeed Insights, enter your URL, and read the field data first, because it shows real visitors.
PageSpeed Insights gives two kinds of result. Field data comes from real Chrome users over the last 28 days, while lab data comes from one Lighthouse run.
Field data is the primary source for TTFB, because it includes redirects that a lab run may miss.
Follow these steps:
- Open PageSpeed Insights and enter the full URL of your home page.
- Read the mobile tab first, then the desktop tab, because the two are rated separately.
- Find TTFB in the field data section. PageSpeed Insights labels it as an experimental metric.
- Repeat the test for 2 or 3 other pages, such as a post, a category page, and the checkout or contact page.
- Note the 75th percentile value for each page, not a single fast load.
A pattern matters more than one number. A TTFB over 1.8 seconds on several different pages is host evidence. A high value on one page only, such as a page with a slow search or a heavy form, points at that page.
If your site has too little traffic for field data, PageSpeed Insights will say so. Use the lab result as a rough check instead, then confirm it with Chrome DevTools on your own connection.
How do you check the server in Chrome DevTools?
Open Chrome DevTools, load the page, and read the Waiting time for the HTML document.
DevTools shows one load in detail, while PageSpeed Insights shows many real visits. Use both, because one slow load could be a one-off, and one fast load could be a cached fluke.
Follow these steps:
- Open the page in Chrome, press F12, and select the Network tab.
- Reload the page and click the first request, which is the HTML document.
- Open the Timing tab for that request.
- Read the row labelled Waiting for server response, which older Chrome versions call Waiting (TTFB).
- Reload 2 or 3 times and note whether the wait stays high or drops after the first load.
A long Waiting time with a short download time means the server took the time, not the network. A short Waiting time with a long download means the page itself is large for the connection.
Compare a logged-out load with a logged-in load. Caching often serves logged-out visitors only, so a logged-in load shows the uncached server time. The difference tells you how much the cache is hiding.
For a related check before you spend money, read how to test hosting speed before buying. It covers the same kind of timing test on a host you do not use yet.
What does WordPress Site Health show about your host?
Site Health shows the server facts WordPress can see, including the PHP version, the database, and failed loopback requests.
Open Tools, then Site Health, in the WordPress admin. The Status tab lists critical issues and recommended improvements. The Info tab lists the raw facts behind them, grouped by section.
Follow these steps:
- Go to Tools, then Site Health, and read the Status tab.
- Open the Info tab and expand the Server section.
- Note the PHP version, the web server, and the PHP memory limit.
- Expand the Database section and note the database type and version.
- Look for a failed loopback request or a late scheduled event, which can point at server limits.
Site Health is built into WordPress, so this check needs no plugin. The WordPress developer reference for WP_Site_Health lists the tests it runs, including PHP version, SQL server version, loopback requests, and HTTPS status.
A loopback failure means WordPress could not make an HTTP request back to itself. That can break background updates and scheduled tasks. It is a server fact, not a theme fact, so take it to the host with the exact message Site Health shows.
How do you check PHP, the database, and memory limits?
Compare the versions in Site Health with the versions WordPress.org recommends, as of November 2026.
WordPress.org recommends PHP 8.3 or greater, MariaDB 10.11 or greater or MySQL 8.4 or greater, and HTTPS support. Apache or Nginx is recommended as the web server.
WordPress still runs on PHP 7.4 and MySQL 5.5.5 in legacy setups, but WordPress.org warns those versions have reached end of life. See the WordPress requirements page for the current list.
PHP support dates matter here. php.net lists PHP 8.2 security support until 31 December 2026, PHP 8.3 until 31 December 2027, and PHP 8.1 as already past end of life on 31 December 2025. A host that cannot offer a supported PHP branch is behind, not just slow.
Use this table to judge what you find in Site Health:
| Item | Recommended, as of November 2026 | Treat as a host problem if |
|---|---|---|
| PHP | 8.3 or greater | Below 8.2, or no upgrade path offered |
| Database | MariaDB 10.11 or greater, or MySQL 8.4 or greater | MySQL 5.7 or older, with no upgrade path |
| Web server | Apache or Nginx | No current server software listed |
| HTTPS | Supported on every install | HTTPS not offered or not working |
If your PHP version is old, ask the host for the upgrade path before you change anything. Read which PHP version to use for WordPress first, because some plugins need a check on staging before a PHP change.
Memory limits are the next fact to read. Site Health shows the PHP memory limit in the Server section.
A limit far below what your plugins need can cause slow admin pages, even when the front end looks fine. If the host will not raise it, that is a plan limit, not a WordPress bug.
How do you test whether caching is hiding a slow host?
Load the same page twice, logged out, and compare the first load with the second, then check the cache headers.
A page cache stores the finished HTML and hands it to the next visitor without running WordPress. That is the fastest path, and it can hide a slow server. An object cache is different: it stores database answers in memory, so WordPress repeats fewer queries on uncached requests.
Follow these steps:
- Open the page in a private window, logged out, and note the load time.
- Reload the same page and note whether it is much faster the second time.
- In DevTools, open the HTML request and read the response headers for a cache status such as HIT or MISS.
- Add a cache-busting query, such as
?nocache=1, and load the page again to see the uncached time. - Repeat the test on a page that must not be cached, such as the cart or a logged-in page.
If the cached load is fast and the uncached load is slow, the host is slow at building pages, and the cache is covering it. That still counts as a host problem for logged-in visitors, search, and checkout, because those requests bypass the page cache.
Cloudflare's Automatic Platform Optimization, known as APO, caches HTML at the edge for WordPress sites. Cloudflare's APO guide says it serves the whole site from the edge network.
APO bypasses the cache for logged-in users, so private content is not shared. To check it, Cloudflare's verify APO page lists three headers: CF-Cache-Status, cf-apo-via, and cf-edge-cache. A HIT value means the edge served the page.
Read Cloudflare cache rules for WordPress before you change cache rules, because a wrong rule can serve a logged-in page to the wrong visitor.
How do you check CPU, PHP workers, and traffic peaks?
Look for slowness that appears only at busy times, because that pattern points at a shared limit rather than a single slow page.
A PHP worker handles one uncached WordPress request at a time. When all workers are busy, the next visitor waits in a queue.
That queue shows up as a high TTFB during busy periods, while quiet periods look fine. Learn the term in what PHP workers are.
Check these signs:
- Pages are fast in the morning and slow in the evening, when traffic is higher.
- The admin area is slow while the cached front end is fast.
- Saving a post, running a search, or loading the cart is much slower than loading a post.
- The host dashboard shows CPU or memory at its limit during the slow periods.
- Errors about resource limits appear in the host panel or in Site Health.
Ask the host for the plan limits in writing: CPU, memory, PHP workers, and any entry-process limit. Then read how to read a hosting plan's CPU and RAM limits and compare the numbers with what your site needs at peak traffic.
Do not guess the limit from one slow afternoon. Record the time, the page, the TTFB, and whether the page was cached. Three dated examples give the host something real to answer, and they give you a baseline to test any fix against.
When should you fix the plan, and when should you move host?
Fix the plan when the cause is a setting you can change; move when the host cannot offer the version, the workers, or the cache your site needs.
Use this decision table after you have the facts above:
| Finding | First step | Move host if |
|---|---|---|
| Old PHP, no upgrade offered | Ask for the upgrade path and a staging test | No supported PHP branch is offered |
| Cache off or misconfigured | Turn on page caching and check the headers | The plan has no server cache and no object cache |
| Workers run out at peak times | Reduce uncached requests and check peak TTFB again | The plan cannot add workers at any price |
| Slow on every page, cached and uncached | Ask the host for the server cause in writing | The host gives no cause and no fix |
| Slow only on one page type | Fix that page, plugin, or query first | Nothing else is slow after the fix |
A move is a last step, not a first guess. Moving a site takes time, and a new host with the same slow plugin or the same missing cache will show the same numbers. Fix the measured cause first, then re-measure.
If you do move, read when to upgrade from shared hosting to judge whether the next step is a better shared plan, a VPS, or managed hosting. A bigger plan helps only when the limit you measured is the one the bigger plan raises.
What mistakes should you avoid?
Do not change hosts on one speed test, and do not change PHP, caching, and the theme in the same hour.
Change one thing at a time and re-measure TTFB after each change. If you change three things at once, you will not know which one helped, and you cannot undo the one that hurt.
Avoid these common errors:
- Judging the host from one lab run, with no field data and no repeat test.
- Comparing a cached load on one host with an uncached load on another.
- Testing while logged in, then wondering why the cache did not help.
- Upgrading PHP on the live site with no staging test for plugins.
- Turning on two cache plugins at once, or a cache plugin plus a host cache, with no header check.
- Ignoring the time of day, so a busy-period queue looks like a permanent fault.
Keep a short log: date, page, TTFB, cached or not, and what you changed. That log is the proof you need when you ask the host for a fix, and it is the proof you need to judge whether a move helped.
Frequently asked questions
What TTFB value means my host is too slow?
A TTFB over 1.8 seconds at the 75th percentile is poor under web.dev guidance, while 0.8 seconds or less is good. Judge the pattern across several pages, not one load. If several pages stay over 1.8 seconds after caching is working, the server is a likely cause and you should ask the host for the reason.
Can a slow host still pass Core Web Vitals?
It can pass on some pages and fail on others, because TTFB is not itself a Core Web Vital. TTFB feeds into Largest Contentful Paint, which is a Core Web Vital with a good limit of 2.5 seconds. A long server wait leaves little time for the rest of the load, so a slow host often pulls LCP over the limit.
Why is my site fast for visitors but slow in the admin area?
The front end is often served from a page cache, while the admin area runs WordPress fresh on every load. Admin slowness can point at PHP workers, database queries, or CPU limits on the host. Check Site Health, compare logged-in and logged-out loads, and ask the host about worker and CPU limits at peak times.
Does a CDN fix a slow host?
A CDN helps with static files and, with HTML caching, can serve cached pages from the edge. It does not fix slow PHP work, slow database queries, or a worker queue on uncached requests. Use the CDN for delivery, and fix the server cause for the requests that still reach the origin.
Should I move host or upgrade my current plan first?
Upgrade or fix the current plan first when the cause is a setting, a PHP version, or a cache you can change. Move when the host cannot offer a supported PHP version, enough PHP workers, or the cache your site needs. Re-measure TTFB after any change so you know whether it helped.
Sources
- Optimize Time to First Byte web.dev
- Time to First Byte (TTFB) web.dev
- Requirements WordPress.org
- Supported Versions PHP
- About Automatic Platform Optimization Cloudflare
- Verify APO works Cloudflare
Is your product missing, or something out of date?
Tell us what to add or correct. We read every note and reply when we need a source. A message does not buy a ranking.
Thanks — we review requests within 3 business days.
