▲ hostfleet /v2
legacy-wp

Web Hosting Speed Comparison 2026: Real TTFB Data, Not Marketing

By Alex Harmon ·

Legacy review from HostFleet’s pre-2026-04-21 era, when we focused on traditional shared and managed WordPress hosting. We have pivoted to AI infrastructure benchmarks (see why). This article remains for reference and people who arrived from search.

Disclosure: If you buy through our links, we may earn a commission at no extra cost to you. Affiliate disclosure.

Every host claims to be “blazing fast.” We run our own edge probes every day and publish the numbers. This page is our current 7-day snapshot of time-to-first-byte (TTFB) across the major affiliate-friendly hosts, updated on 2026-04-20. Full collection details live on our methodology page.

How we measure

We request each host’s public marketing URL on a recurring schedule from a single VPS location, then log time_starttransfer from curl. We aggregate the last seven days and report median and mean TTFB. We treat 2xx and 3xx responses as valid samples and flag 4xx/5xx separately. Full monitoring details, caveats, and collection intervals are documented in our methodology.

This is not a plan-on-plan lab benchmark. It is an honest baseline: if a host’s own marketing site is slow, their customer sites usually are not magically faster.

Current 7-day snapshot (2026-04-14 to 2026-04-20)

  • Cloudways — 188 ms median TTFB (n=178 valid 302 responses to the regional site, with 9 separate 403 responses flagged outside the main median). Fastest host in our current valid-response dataset.

  • Hostinger — 203 ms median TTFB (n=187, all 200). Effectively tied with Cloudways on median, slightly wider tail.

  • ScalaHosting — 248 ms median TTFB (n=187, all 200). Solid mid-pack result for a managed VPS marketing site.

  • SiteGround — 529 ms median TTFB (n=187). Slower in our measurement, likely partly because our probes get redirected to the German storefront.

  • Bluehost — not measurable. Our probes consistently return HTTP 403. We flag this as a data gap and have done so since we started publishing these numbers.

Reading these numbers honestly

  • Median matters more than average on a small window. One slow response can blow up the mean. We report both so you can see the spread.

  • Your server TTFB will differ. Your site, plan tier, region, cache setup and plugin stack are not on this chart. Take our numbers as a speed floor, not a ceiling.

  • A fast marketing site is not a fast plan. Hosts pour money into their own frontend. The gap between the homepage and your WordPress site can be significant.

What this does and does not tell you

It tells you which hosts are at least capable of serving a basic page quickly from a normal internet location. It does not tell you which shared plan will survive a Reddit hug-of-death or which managed WordPress plan gives you the best staging workflow.

For plan-level recommendations, pair these numbers with our written reviews: Hostinger, Cloudways, ScalaHosting, SiteGround, and Bluehost.

Methodology notes

  • Single-region probe. Global CDN performance is not captured.

  • Public marketing URLs, not customer-plan test sites.

  • 7-day rolling window. We re-run this aggregation weekly.

  • We do not cherry-pick individual fast days.

If you want us to test a specific host that is not on this list, let us know. We add providers when we have enough data to publish honestly.

By Alex Harmon, Lead Hosting Analyst — HostFleet. Last updated 2026-04-20.