To test web hosting speed fairly, put the same test website on each host, measure server response time and page-load metrics from locations near your real visitors, repeat the tests several times over days, and compare medians rather than single runs. Published “fastest host” rankings are only useful if they explain this method; most do not. This guide shows you how to run the test yourself.
What actually makes a host fast
- Server response time. How long the server takes to start sending the page. It depends on CPU, memory, storage, web server software and how busy the server is.
- Storage. NVMe drives read and write much faster than older SATA SSDs or hard drives, which matters for databases.
- Web server and caching. LiteSpeed with its cache plugin, or Nginx with page caching, can serve WordPress pages without running PHP for every visit.
- PHP version. Newer supported PHP versions generally run WordPress faster than old ones.
- Distance. Every request travels to the server and back. A server far from your visitors adds delay to each round trip.
- Content delivery network. A CDN such as Cloudflare serves cached files from locations near each visitor.
- Your own site. Large images, heavy themes and many plugins slow a page down on any host.
The metrics to measure
Time to First Byte (TTFB)
The time from the request to the first byte of the response. It is the clearest measure of the host itself, because it is mostly server and network time.
Core Web Vitals
Google’s Core Web Vitals measure what visitors experience. According to web.dev’s article How the Core Web Vitals metrics thresholds were defined (last updated 2025), a page is “good” at the 75th percentile when:
- Largest Contentful Paint (LCP) is 2.5 seconds or less.
- Interaction to Next Paint (INP) is 200 milliseconds or less.
- Cumulative Layout Shift (CLS) is 0.1 or less.
LCP is affected by hosting through TTFB. INP and CLS depend mostly on your theme, scripts and page layout, so they say more about the site than the host.
Lab data vs field data
Lab tests load a page under controlled conditions from a chosen location. They are repeatable and good for comparing hosts. Field data comes from real visitors’ browsers and shows what people actually experience, but it needs traffic to build up and mixes hosting with everything else on the page. Use lab tests to compare hosts, and field data to confirm the result after you move.
A practical method
- Build one test site. Use the same WordPress version, theme, plugins, content and images on every host. A copy of your real site is ideal.
- Match the settings. Use the same PHP version and the same caching approach on each host, or test both with and without caching and label the results.
- Choose test locations close to where your visitors are, and add one distant location for comparison.
- Measure TTFB with a tool such as WebPageTest, Chrome DevTools or a command-line request that reports time to first byte.
- Measure full page load with PageSpeed Insights or WebPageTest, recording LCP, CLS and total page weight.
- Repeat. Run each test several times at different times of day over at least a few days, and use the median for each host.
- Separate cold and warm cache. The first request after a cache clear is slower than later ones. Record both.
- Test the uncached path. Log in to WordPress or load a search results page to see how fast the server is when it has to run PHP and query the database.
- Write down the date, plan and settings for every result so you or anyone else can repeat the test.
Common mistakes
- Testing a blank default site, which hides differences in database and PHP performance.
- Comparing a cached page on one host with an uncached page on another.
- Running a single test and treating it as the result.
- Testing from a location far from your real visitors.
- Comparing different plan sizes, such as an entry plan against a premium one.
- Loading a shared server with stress tests, which may breach its acceptable use policy.
Improving speed on the host you have
- Turn on page caching and compress images before upload.
- Use a supported, recent PHP version.
- Remove unused plugins and heavy page-builder elements.
- Put a CDN in front of the site.
- Choose a server location near your visitors, or a larger plan if you hit resource limits.
How Tremhost hosting is set up
Tremhost web hosting runs on cPanel with LiteSpeed web servers and NVMe storage, on servers in Europe and the USA. The Big Mike plan adds LiteSpeed Enterprise and Cloudflare Pro for the main domain. Tremhost is a Cloudflare partner, and Cloudflare has a data centre in Harare. Armor Guard, which is Cloudflare Pro configured by Tremhost, is $9 a month for one site, and you can try Cloudflare Pro free for 15 days without a card. Run the method above on your own site to see how it performs for your visitors.
Frequently asked questions
What is a good TTFB?
Lower is better, and it varies with distance and caching. Compare TTFB between hosts from the same location rather than against a single target.
Can I trust “fastest host” rankings?
Only if they publish the test site, locations, dates, number of runs and settings. Without that, the results cannot be checked or repeated.
Does a CDN make hosting speed irrelevant?
No. A CDN speeds up cached files, but uncached pages, logged-in users and checkouts still depend on the server.
Next step
Compare plans on the Tremhost web hosting page, or ask on WhatsApp at +263 73 563 9917 (Monday to Friday, 08:00 to 17:00 CAT) for help setting up a test site.
