Plugins Plugin hub Backup SEO Caching Security Forms WooCommerce Images Best backup plugins
Hosting Hosting directory Shared Managed Cloud India
Themes & builders Compare Rank Math vs Yoast
Tools All tools Hosting finder Plugin stack builder Cost calculator Salt generator Theme detector Who is hosting this? Speed test Headline analyzer robots.txt generator Redirect generator Uptime calculator Blog name generator
Best lists Best backup plugins Deals
Guides All guides Speed up WordPress Choose hosting Back up WordPress
How we test

Hosting · Guide

How to Test a Host's Speed Before You Buy

Short answer

To test a host's speed before you buy, build the same small WordPress test site on each shortlist host, then run PageSpeed Insights, Lighthouse and WebPageTest on the same pages with the same settings. Compare LCP, INP, CLS and TTFB across at least 3 runs each, and check the renewal price and refund window before you pay.

Key takeaways

  • A good LCP is 2.5 seconds or less, a good INP is 200 milliseconds or less, and a good CLS is 0.1 or less, all at the 75th percentile.
  • PageSpeed Insights shows real visitor data for the past 28 days plus one Lighthouse lab run, so read the field data first.
  • A Lighthouse audit in Chrome DevTools takes 30 to 60 seconds and works on staging sites and pages behind a login.
  • Run each test at least 3 times with the same location, device and connection, then compare the middle result.
  • WordPress.org recommends PHP 8.3 or greater, MariaDB 10.11 or greater or MySQL 8.4 or greater, and HTTPS support as of November 2026.

What does testing a host's speed before you buy actually mean?

Testing a host's speed before you buy means measuring the same test site on each host with the same tools and settings, then comparing the numbers.

A host speed test is not a single score from one run. One run can be fast by luck or slow because of a busy minute on a shared server. You need repeated runs, the same pages, and the same test settings on every host you compare.

A heavy theme, 20 plugins, and uncompressed images can make any host look slow, so a light test site shows what the server itself adds. If you want help narrowing the shortlist first, read how to choose WordPress hosting.

What do you need before you start testing?

You need a shortlist of 2 or 3 hosts, a test site you can copy to each one, and a note of the exact test settings you will use.

Gather these items first:

  • A shortlist of 2 or 3 hosts with a trial, a money-back window, or a staging plan you can cancel.
  • The same WordPress install on each host: the same theme, the same 3 or 4 plugins, and the same 3 pages.
  • A home page, a blog post, and a page with one image, so you test more than the lightest page.
  • A notebook or sheet with columns for host, tool, location, device, run number, LCP, INP, CLS, and TTFB.
  • A clock and a plan to run the tests at a similar time of day on each host.

Set aside about 2 hours. Building the test site takes most of that time.

Which numbers should you judge a host by?

Judge a host by LCP, INP, CLS, and TTFB, because those 4 numbers cover loading, response, stability, and server delay.

LCP, or Largest Contentful Paint, marks when the largest visible part of a page finishes loading. INP, or Interaction to Next Paint, measures how fast a page responds when a visitor clicks or taps. CLS, or Cumulative Layout Shift, measures how much visible content moves while a page loads.

TTFB, or Time to First Byte, is the time from the start of a page request until the first byte of the response arrives. It is a diagnostic number, not a Core Web Vital, but it is the closest public number to raw server speed.

Web.dev's thresholds page sets the good and poor lines for the 3 Core Web Vitals at the 75th percentile of page views. That means at least 75% of visits must meet the good line for a page to pass.

MetricWhat it measuresGoodPoor
LCPMain content loading2.5 seconds or lessOver 4.0 seconds
INPResponse to clicks and taps200 ms or lessOver 500 ms
CLSLayout stability0.1 or lessOver 0.25
TTFBServer response start0.8 seconds or lessOver 1.8 seconds

The TTFB lines come from web.dev's TTFB guide, which calls 0.8 seconds or less a rough guide for most sites. Learn more about one of these numbers in what INP is.

How do you test a host with PageSpeed Insights?

Test a host with PageSpeed Insights by entering a public page URL at pagespeed.web.dev and clicking Analyze, then reading the field data at the top before the lab score.

PageSpeed Insights is free and needs no account. Google's PageSpeed Insights guide says the tool shows both lab data and field data about a page. Lab data comes from one Lighthouse run. Field data comes from the Chrome User Experience Report, known as CrUX, and covers real Chrome users over the previous 28 days.

Follow these steps for each host:

  1. Open PageSpeed Insights in your browser.
  2. Paste the full URL of the test home page, including https://.
  3. Click Analyze and wait for both sections to load.
  4. Read the top section, headed “Discover what your real users are experiencing.”
  5. Note LCP, INP, CLS, and TTFB, plus whether the Core Web Vitals assessment says Passed or Failed.
  6. Scroll to the lab section and note the performance score from 0 to 100.
  7. Switch between the Mobile and Desktop tabs and note both sets of numbers.
  8. Repeat the run 2 more times and keep the middle result for each number.

A new test site has no field data, because CrUX only covers pages with enough real traffic. On a new site, use the lab numbers and TTFB.

How do you test a host with Lighthouse in Chrome?

Test a host with Lighthouse in Chrome by opening the page, opening DevTools, clicking the Lighthouse tab, and clicking Analyze page load.

Lighthouse is an open-source audit tool from Google. Chrome's Lighthouse guide says you can run it in Chrome DevTools, from the command line, as a Node module, or through PageSpeed Insights. It audits performance, accessibility, SEO, and more, and it gives a report after 30 to 60 seconds.

Follow these steps:

  1. Open the test page in Chrome, in a new incognito window.
  2. Right-click the page and select Inspect, or press Ctrl+Shift+I on Windows or Cmd+Option+I on Mac.
  3. Click the Lighthouse tab at the top of DevTools.
  4. Leave the categories enabled and select Mobile as the device.
  5. Click Analyze page load, then Run audit when asked.
  6. Wait 30 to 60 seconds for the report.
  7. Note the performance score and the LCP, INP, CLS, and TTFB values.
  8. Run the audit 2 more times and keep the middle result.

Lighthouse in DevTools has one clear edge over PageSpeed Insights: it works on staging sites and pages behind a login. Chrome's DevTools guide warns that Lighthouse is affected by your own machine, open tabs, and extensions, so close other tabs and use incognito mode, and run every host from the same computer.

How do you test a host with WebPageTest?

Test a host with WebPageTest by entering the page URL, picking a test location near your visitors, picking Chrome as the browser, and clicking Start Test.

WebPageTest runs a real browser from a real location, which makes it useful for checking how distance affects speed. WebPageTest's getting started guide lists the same order: enter the page URL, select a location, select a browser, then submit the test. It recommends Chrome for first tests because it reports the most detail.

Follow these steps:

  1. Open WebPageTest and enter the test page URL.
  2. Pick the test location closest to your main visitors.
  3. Pick Chrome as the browser and Mobile or Desktop to match your audience.
  4. Set the number of runs to 3 if the option is shown.
  5. Click Start Test and wait for the results page.
  6. Note LCP, CLS, TTFB, and the fully loaded time on the first view.
  7. Open the waterfall chart and look at the first row, which is the HTML page itself.
  8. Repeat the test from a second location if your visitors live on more than one continent.

The waterfall is the most useful part of a WebPageTest report. A long first row points at the server, while long image or script rows point at the page. Keep the location, browser, and connection the same across hosts.

How do you build a fair test site on a trial or staging plan?

Build a fair test site by installing the same WordPress version, theme, plugins, and pages on each host, then testing before you add anything else.

A fair test site is small and identical everywhere. Use a default theme such as Twenty Twenty-Five, or the same light theme on each host. Add only the plugins the site truly needs, and use the same 3 pages with the same text and the same 1 or 2 images.

Follow these steps on each host:

  1. Sign up for the trial or money-back plan and note the date the refund window ends.
  2. Install WordPress from the host panel and note the PHP version shown.
  3. Install the same theme and the same plugins on each host.
  4. Create the same home page, blog post, and image page on each host.
  5. Turn off any host caching or CDN option that is off by default on the other hosts, or turn the same option on everywhere.
  6. Test the site while logged out, so the admin bar does not change the numbers.

Many managed hosts include a one-click staging copy. Use it to test a change without touching the live site, and make a backup first, as explained in how to back up WordPress.

How do you read TTFB and tell server speed from page weight?

Read TTFB first to judge the server, then read LCP and the waterfall to judge the page, because TTFB measures the wait before the browser can start work.

A low TTFB with a slow LCP usually means heavy images, fonts, or scripts are slowing the page after the server has answered. That split tells you where to fix things.

Use this short checklist when you read a report:

  • TTFB over 1.8 seconds on a light test page points at the server or the route to it.
  • TTFB under 0.8 seconds with LCP over 4 seconds points at page weight, not the server.
  • A first waterfall row that is much longer than the other rows points at the server.
  • The same page tested from a far location will show a higher TTFB, so compare hosts from the same location.

Check the PHP version on each host too. Go to Tools, then Site Health, then Info, then expand Server and read the PHP version row.

WordPress.org's requirements page recommends PHP 8.3 or greater, MariaDB 10.11 or greater or MySQL 8.4 or greater, and HTTPS support as of November 2026. If you want to fix page weight after the host choice, read how to speed up WordPress.

How do you compare hosts fairly side by side?

Compare hosts fairly by using the same pages, tools, locations, and run count on each host, then comparing the middle result from each set of runs.

Write the numbers in a table as you go. Do not trust a single best run. The middle of 3 runs is a steadier guide than the fastest run, which may have used a warm cache.

Use this order for the comparison:

  1. Run PageSpeed Insights on the home page of each host, 3 times, on Mobile.
  2. Run Lighthouse in Chrome on the same page of each host, 3 times, on Mobile.
  3. Run WebPageTest on the same page of each host, 3 times, from the same location.
  4. Repeat steps 1 to 3 on the blog post page.
  5. Write the middle LCP, INP, CLS, and TTFB for each host in one table.
  6. Mark a host as faster only when it wins on most numbers across most pages.

A host that wins by 0.1 seconds on one page and loses by 1 second on the next has not won. Look for a clear gap that repeats. For a worked example, read SiteGround vs Hostinger speed.

What should you check in the price and refund terms?

Check the renewal price, the term length, and the refund window on the host's own pricing page, because the first-term price is not the price you will pay in year 2.

Intro prices apply to the first term only. A 48-month term locks in a low monthly rate for 4 years, while a 12-month term renews sooner at a higher rate. Read the renewal price on the plan card before you pay, and write it next to the intro price.

The table below shows renewal prices as of November 2026. Prices change with promotions, so confirm the live price on each host's pricing page before you buy.

HostEntry plan renewal as of November 2026Money-back window
SiteGroundStartUp renews at $17.99 per month30 days on shared hosting
HostingerPremium renews at $10.99 per month30 days
KinstaSingle-site plans start at $35 per month30 days

SiteGround's StartUp plan renews at $17.99 per month after a first term priced from $1.99 to $2.99 per month, as shown on SiteGround's WordPress hosting page.

Hostinger's Premium plan renews at $10.99 per month after an intro price of $2.49 to $2.99 per month on a 48-month term, as shown on Hostinger's WordPress hosting page.

Kinsta's single-site plans start at $35 per month, or $30 per month billed yearly, as shown on Kinsta's pricing page.

Also check the visit, storage, and backup limits on the plan card. A plan with a 10,000-visit limit or 10 GB of storage may fit a new blog but not a shop.

How do you know the test worked?

You know the test worked when you have 3 runs per tool per host, the same settings everywhere, and a clear winner or a clear tie you can explain.

Check your notes against this list before you decide:

  • Each host has at least 3 PageSpeed Insights runs, 3 Lighthouse runs, and 3 WebPageTest runs.
  • Every run used the same pages, location, device, and connection.
  • You recorded the middle result for LCP, INP, CLS, and TTFB on each host.
  • You checked the PHP version and the renewal price on each host.
  • You noted the refund deadline for the host you picked.

If one host is faster on most numbers and its renewal price fits your budget, that host is the pick. If the numbers are close, pick the host with the better renewal price and clearer limits.

What mistakes should you avoid?

Avoid testing different pages, different settings, or a single run, because any of those mistakes can make the slower host look faster.

The most common mistakes are easy to avoid:

  • Testing only the home page, which is often the lightest page on a site.
  • Testing one host on Mobile and the next on Desktop.
  • Testing one host from a near location and the next from a far location.
  • Comparing a cached run with a cold run and calling the gap a host difference.
  • Trusting a host's own speed badge or a demo page you did not build.
  • Judging a host by the 0 to 100 Lighthouse score alone, while ignoring LCP, INP, CLS, and TTFB.
  • Buying a long term for the intro price without reading the renewal price.

Run the tests again after you move the real site, using the same tools and settings.

Frequently asked questions

Can I test a host's speed without buying a plan first?

Yes, in most cases. Use a money-back window, a free trial, or a staging plan, and build the same small test site on each host you compare. PageSpeed Insights and WebPageTest work on any public URL, and Lighthouse in Chrome works on staging sites and pages behind a login. Cancel inside the refund window if the host is not the right fit.

How many times should I run each speed test?

Run each test at least 3 times with the same location, device, and connection, then use the middle result. A single run can be fast or slow by chance, especially on shared hosting. Three runs show whether a result repeats. If the 3 runs differ widely, run 2 more and use the middle of the 5.

Why do PageSpeed Insights and GTmetrix give different scores?

They test in different ways. PageSpeed Insights shows real visitor data from the Chrome User Experience Report plus one Lighthouse lab run on Google servers. GTmetrix runs Lighthouse in a real Chrome browser from a location you pick, with its own connection settings. Different locations, devices, and cache states change the numbers, so compare results only when the settings match.

Is TTFB the same as page load time?

No. TTFB measures only the wait from the start of a request until the first byte of the response arrives. Page load time covers the whole page, including images, fonts, and scripts. A low TTFB with a slow LCP usually means the page itself is heavy. A high TTFB on a light test page usually points at the server or the route to it.

What is a good LCP, INP, and CLS score for a WordPress site?

A good LCP is 2.5 seconds or less, a good INP is 200 milliseconds or less, and a good CLS is 0.1 or less, all measured at the 75th percentile of page views. Poor lines are over 4.0 seconds for LCP, over 500 milliseconds for INP, and over 0.25 for CLS. A page passes only when all 3 numbers are in the good range.

Should I pick the cheapest host if its speed numbers are close to a dearer host?

Only after you compare the renewal price, not the intro price. Check the visit and storage limits, the backup policy, and the refund window too. If the speed gap is small and the cheaper host renews at a much lower price with limits that fit your site, it is often the better pick. If the cheaper host has tight limits, the dearer host may cost less over 2 years.

Sources

  1. How the Core Web Vitals metrics thresholds were defined web.dev
  2. Time to First Byte (TTFB) web.dev
  3. About PageSpeed Insights Google for Developers
  4. Introduction to Lighthouse Chrome for Developers
  5. Lighthouse: Optimize your website Chrome for Developers
  6. Requirements WordPress.org
  7. Welcome to WebPageTest WebPageTest Documentation
  8. PageSpeed Insights Google

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.

About the editors

WPLegible editors

Researched from the primary sources listed above and checked against our editorial rules. About the site · How we test