Methodology
How We Test Hostinger
Our methodology: what we run, what we measure, how long we observe it, and what we refuse to claim.
Last updated
We do not run a laboratory. We run real sites on the plans we review, watch what happens, and report what the published limits mean in practice. Here is exactly what that involves — including the parts we cannot measure well and do not pretend to.
What we run
For every plan tier we review, we deploy a standard test site rather than a synthetic benchmark. The stack is deliberately ordinary, because an ordinary stack is what most readers will build:
- A default WordPress install on the current major version
- One page builder, one forms plugin, one caching plugin — nothing exotic
- A WooCommerce variant on the plans we recommend for stores
- A plain static HTML site as a control, so we can separate host performance from application overhead
- A Node.js application on the tiers that support it, to check that the advertised support is real
What we measure
| What | How | Why it matters |
|---|---|---|
| Uptime | An external monitor polling the test site at a fixed interval from two regions | Hostinger publishes a 99.9% commitment, which still permits roughly 43 minutes of downtime a month |
| Time to first byte | Repeated requests against a cached and an uncached page | TTFB is where cheap hosting shows its limits before anything else does |
| Resource ceiling | Loading the site until we hit PHP worker, memory or inode limits | This is the number that predicts when you will see a 503 page |
| Support quality | Timed live-chat requests on real, specific problems | Response time is published; whether the answer actually solves the problem is not |
| Control panel tasks | Timing common jobs: pointing a domain, issuing SSL, creating a mailbox, restoring a backup | These are the tasks beginners get stuck on |
What we do not do
We are explicit about the limits of our testing, because a review that overstates its rigour is worse than one that admits its scope.
- We do not publish precise benchmark scores. A single TTFB number from one location on one day is not a meaningful comparison between hosts, and presenting it as one misleads readers. We report ranges and the conditions we observed.
- We do not test every data centre. Performance varies by region. If you are serving an audience far from the data centre you choose, your results will differ from ours.
- We do not invent customer quotes. Every testimonial on this site is either sourced from a public review platform with a link, or it does not appear.
- We do not claim long-run statistics from short-run observation. If we watched a site for 30 days, we say 30 days.
How we score
Plan scores are out of 5 and weighted toward the things that change your bill and your workload, not the length of the feature list:
| Criterion | Weight |
|---|---|
| Total cost of ownership (intro plus renewal) | 30% |
| Resource allocation relative to price | 25% |
| Reliability and backup policy | 20% |
| Ease of use and control panel | 15% |
| Support quality | 10% |
That weighting is why a plan with a spectacular introductory price and a steep renewal rate scores lower here than it does on sites that only quote year one. It is also why Unlimited consistently outscores Premium: not because it is more expensive, but because the extra dollar buys disproportionately more.
How we verify prices
Prices, term lengths and renewal rates are recorded directly from hostinger.com and from Hostinger's own published parameters documentation, with the verification date stored alongside the data. The current verification date is 2026-09-17.
Where Hostinger's marketing pages and its technical documentation disagree — the most notable case being cloud plans marketed as "unlimited websites" while the parameters table lists a 100-site cap — we publish both and say which is which.