How Tech Hosting Guide tests and scores hosting
Our complete testing methodology, scoring weights, conflict-of-interest rules and correction policy. Published so you can check our working.
Last reviewed 2026-08-20
This page describes exactly how Tech Hosting Guide tests and scores web hosting providers. It exists so that anyone reading a review on this site can check our working, reproduce our tests, and judge for themselves whether our conclusions are reasonable.
We publish this because the single biggest problem with hosting reviews online is that most of them are not reviews. They are affiliate pages with a star rating attached. A review that cannot tell you how it reached its score is not telling you anything.
The six criteria and their weights
| Criterion | Weight | What we measure |
|---|---|---|
| Performance | 25% | Server response time, time to first byte, and page load under a controlled synthetic load, measured from multiple geographic locations against an identical WordPress installation. |
| Uptime & reliability | 20% | Continuous external monitoring at one-minute intervals across the full measurement window, plus the provider's published SLA and what it actually pays out when breached. |
| Support quality | 20% | Timed, scripted support tickets and live chat sessions submitted at different hours and on different days, scored on first-response time, resolution rate, and whether the answer was correct. |
| Pricing & value | 15% | Introductory price, renewal price, the contract term required to reach the advertised rate, what is genuinely included versus sold as an add-on, and the refund terms. |
| Ease of use | 10% | Signup friction, control panel clarity, migration tooling, one-click installers, and how many steps it takes to get a working site online from a cold start. |
| Features & flexibility | 10% | PHP and database versions available, staging environments, SSH and Git access, backup frequency and restore cost, email provision, and resource limits as actually enforced. |
The test environment
Every provider is tested on the plan we believe a typical small-business or personal customer would actually buy — usually the mid-tier shared hosting plan, not the cheapest entry plan and not a managed or VPS product. We note the exact plan on each review page.
Each account receives an identical build: a clean WordPress installation, the same lightweight theme, the same twenty seed pages and fifty seed posts, the same media library, and the same caching configuration where the host allows us to control it. Where a host forces its own caching layer, we say so, because it affects the numbers.
Performance measurement
Response time and time to first byte are sampled at regular intervals over the full measurement window from several geographic locations, so that a host with one fast region and three slow ones cannot hide behind an average taken from a single city.
Load testing ramps concurrent virtual users against the same page set on every host. We record the point at which response time degrades and the point at which errors begin, rather than reporting a single headline number.
We publish the measurement window dates on every review. A performance figure without a date is worthless, because hosting infrastructure changes.
Uptime measurement
Uptime is measured by external monitoring at one-minute intervals for the duration of the window. We count any non-200 response or timeout as downtime. We report the raw percentage, the number of distinct incidents, and the longest single outage, because a single four-hour outage and forty short blips produce similar percentages and very different experiences.
Support testing
We submit a fixed set of support requests to every provider: one simple factual question, one genuine technical problem, and one billing question. These are sent at different times of day and on different days of the week, including at least one outside standard business hours in the provider's home time zone.
We score first-response time, whether the first response actually answered the question, and how many exchanges the resolution required. We do not identify ourselves as reviewers.
Pricing analysis
Every pricing table on this site shows the introductory price, the contract term required to obtain it, and the renewal price. Where a provider's advertised headline price requires a multi-year prepayment, we say so next to the number.
Hosting prices change frequently and often vary by region and by promotion. We date-stamp every pricing table and link to the provider's official pricing page as the authoritative source. If you find a discrepancy, please tell us and we will correct it.
How affiliate relationships are handled
Tech Hosting Guide earns revenue from affiliate commissions. We think readers are entitled to know exactly how that works and exactly what it does not affect.
- Providers are selected for review based on market share and reader demand, not on commission rates. Providers with no affiliate programme are reviewed on the same terms as providers with one.
- Scores are calculated from the measured results and the published weights. Commission rates are not an input to the score and cannot be.
- The person who writes and scores a review does not have access to commission data for the provider being reviewed.
- No provider is given advance sight of a review, a score, or a draft, and no provider has the right to request changes.
- Every affiliate link on this site is marked with
rel="nofollow sponsored"and carries a visible disclosure.
Update and correction policy
Reviews are re-tested at least annually, and sooner when a provider makes a material change to pricing, infrastructure, or plan limits. Every page shows the date it was last updated.
When we publish something inaccurate we correct it on the page, add a dated note explaining what changed, and do not quietly overwrite the record. If you believe something on this site is wrong, email contact@techhostingguide.site and we will respond.