HOW WE MEASURE HOSTING SPEED & PERFORMANCE
Choosing a web host often starts with one question: how fast is it? At HostList, performance is one of four equally-weighted inputs to HRI (HostList Ranking Index) (HRI™). We do not rely on vendor marketing claims. We use controlled benchmarks where available, real-user Core Web Vitals where not, and clearly label weaker estimates when data is missing.
Hosting speed vs site speed
Hosting speed is how quickly a server responds when a visitor requests a page, mainly time to first byte (TTFB) and network latency. It depends on server hardware, data-centre location, PHP/runtime configuration, and database performance.
Site speed is how fast a full page loads in a browser. A heavy theme, unoptimized images, or plugin bloat can slow a site even on powerful hosting. Conversely, a well-optimized site can load quickly on average infrastructure.
HRI focuses on signals that reflect hosting infrastructure. When we use CrUX field data, it comes from the provider's public website. That is useful, but not identical to how your site would perform on their platform. Controlled benchmarks (identical test site on the host's real plan) are the gold standard. We are manually deploying identical test website instances across the top 200 hosts by HRI rank. See also: TTFB, Core Web Vitals.
The HRI performance ladder
Each host gets the strongest available signal. We never mix tiers; a benchmark always wins over CrUX or estimates.
An identical WordPress test site load-tested on the host's real plan (e.g. ReviewSignal). Metrics: median response time under load, uptime % during the test window, error rate, Top Tier award. Shown on profiles as "Benchmarked".
Google CrUX p75 field data (LCP, INP, CLS, TTFB) on the provider's public website origin. Refreshed automatically. Discarded after 120 days. Does not measure customer sites on the platform. We are manually deploying identical test website instances across the top 200 hosts by HRI rank.
Editorial infrastructure tier (CDN stack, caching layer, known platform) when no benchmark or CrUX data exists.
Neutral 11/25 starting score, not a penalty. Rises as data is collected or operational signals are verified.
Full formulas and weightings: HRI™ methodology.
What we measure (and what we don't)
We measure: response time under load (benchmarks), uptime during benchmark windows, error rates under load, CrUX p75 on provider origins, public status-page presence, and infrastructure tier as a last resort.
We don't measure: your theme, image optimization, plugin choices, continuous multi-year uptime for every host, or every plan tier from every provider. Shared plans are not compared directly to dedicated servers; tier context matters when reading rankings.
Test your own hosting speed
Every host claims to be fast. The only speed that matters is what your visitors experience on your site.
- Google PageSpeed Insights: check TTFB in the lab report; CrUX section shows real-user data if your site has enough traffic.
- WebPageTest: TTFB and waterfall from multiple regions.
- Bitcatcha: quick global latency ping from key regions.
- UptimeRobot: free HTTP monitor at 5-minute intervals for a 30-day reliability snapshot.
- Loader.io: simulate a traffic spike; even a few hundred concurrent requests reveals whether the server buckles.
Frequently asked
What is the difference between hosting speed and site speed?
Hosting speed measures how quickly a server responds to a request, mainly time to first byte (TTFB) and latency. Site speed measures how fast a full page loads in a browser, which depends on themes, images, plugins, and caching as much as the server. A heavy WordPress site can be slow on a fast host; a well-optimized site can feel quick on average hosting.
What does "Benchmarked" mean on a host profile?
Benchmarked means an independent controlled test ran an identical site on the host's real plan under load, for example ReviewSignal's WordPress hosting benchmarks. It measures response time under load, uptime during the test window, and error rate. This is the strongest performance signal in HRI because it isolates server infrastructure from marketing-site design.
Can I trust CrUX data for comparing hosts?
CrUX field data reflects real Chrome users visiting the provider's public website, usually their marketing homepage. It is useful when no benchmark exists, but it mixes server performance with how that specific site is built. It does not tell you how fast your site would run on their platform. We are manually deploying identical test website instances across the top 200 hosts by HRI rank. See the FAQ below for rollout status.
Are you deploying identical test sites on hosts?
Yes, in progress. We are manually deploying identical test website instances across the top 200 hosts by HRI rank. Benchmark data is published host by host as each deployment is validated and load-tested on the host's real plan. Profiles move from Measured or Estimated to Benchmarked as results go live. Until benchmark data exists for a host, Performance falls back to CrUX field data on the provider's public website (Measured) or an infrastructure tier estimate (Estimated).
Does HostList run continuous uptime monitoring on every host?
No. Uptime figures in HRI come from controlled benchmark test windows and status-page signals where detected. We do not publish multi-year UptimeRobot dashboards for every provider. For ongoing reliability, we recommend monitoring your own site or checking the host's public status page.
How can I test my own hosting speed?
Start with Google PageSpeed Insights or WebPageTest and check TTFB. Use Bitcatcha or WebPageTest regional runs for latency from your audience's geography. Set up UptimeRobot for a 30-day reliability snapshot. For load handling, try Loader.io with a few hundred concurrent requests. Always test your actual site, not the host's homepage.
Founder, HostList.io · Co-founder, Seahawk Media · Host of the WP Legends podcast.