How TechRadar Pro tests web hosting providers

I take a qualitative and quantitative approach when it comes to finding the best web hosting providers. For some things, metrics are not important. You can always buy more resources but you cannot just magic a fix to a server issue when you need to. In these cases, support is more important than speed. In other cases, such as web stores, speed can mean losing sales. So, when we do our tests we focus on the area that makes the most sense to test. That can mean setting up an unmanaged VPS and seeing how easy it is to manage yourself. Creating Flask-Python apps that require a custom server environment and migrating it from one host to another. It can also mean creating a complex web store or a simple WordPress blog.

These tests give us the knowledge to recommend the best web hosting providers for a variety of use cases from the best free web hosting to the best WordPress hosting.

Support

If all you want is a server, support isn't a purchasing factor but most people are buying more than just physical parts. Migration, custom server environments, security updates, and performance optimization are the things that are the top priorities and what is making the price tag significantly larger.

That's why we do more than just ask a few basic questions to support. We build web applications and ask support to help us get our server environment up and running. In doing so we've discovered which hosts go the extra mile in solving issues, which ones just want to close a ticket as soon as possible regardless of if the problem still exists, and hosts that have support workers with sub-par technical knowledge.

That's why you can trust our recommendations for the best VPS hosting providers, best cheap VPS hosting, and best managed web hosting.

Comparing features

On top of hardware and support you're also paying for features. So, these are things we look at too. Who has a larger CDN network? Who will give you the most email accounts? We look at these features among others when making recommendations but not only that, along with testing them, we also help put things in perspective.

We tell you when it's not a big deal if hosts are lacking features too. If you don't need daily back ups and weekly back ups will do and when getting less storage can actually mean more due to the small print. This can help you save money and get the best value for money when shopping for things like the best cheap WordPress hosting.

Web hosting pricing

Pricing is one of the most confusing aspects of hosting. There's one price, and then a discount price, and then a renewal price. We aim to present prices as clearly as possible and we only take into account the total cost of ownership when we recommend a host.

If we recommend something as free, it will always be free and not just a trial. If we say something is cheap it is cheap the entire life of the service or we will recommend where to go next once your discount period is over to keep your hosting cheap. We don't fall for bait and switch tactics and we point out when hosts are trying to pull a fast one.

Ease of use

When we're testing a host we always check out what the provider does to help new users: the first emails they send, the design of the account control panel, how easy it is to find your way to the hosting dashboard and all the core site-building tools.

We explore the options in more detail, too. Is there a way to install WordPress? Does it support other apps? Is there a template-based Site Builder, or other useful extras? If the plan doesn't include a standard dashboard like cPanel, what do you get, and how well does that solution work?

Not everyone has the years of experience we have and sometimes it's easy to forget and assume knowledge so we often get beginners onboard too to make their first website. This way we really know what it's like for beginners and are very confident in our recommendations for this category across all our guides.

Trustworthiness

Web hosting and email are vital services for personal and especially business users, so it's important to choose a provider you can trust. Assessing this is difficult, but we look for clues in several areas.

Does the company present its products in an honest, clear and transparent way, for instance? Or does it make over-the-top promises, give you minimal details on the products, quote headline prices but not how long you must sign up to get them, or hide away important details in the small print?

A good host should have a well-maintained web presence, professionally presented and fully up to date. If news pages, blogs and social media accounts haven't been updated in an age, that doesn't suggest a company that's actively working to improve its services. And broken links, certificate errors or other fundamental flaws indicate a serious lack of resources. What other corners might the host be cutting to save a little cash?

For us, though, the strongest indicators of an untrustworthy host appear after we've signed up and are testing the service.

For instance, sometimes our account might not be immediately usable due to some tech issue, so we raise a support ticket. A good host should respond quickly, either fix the issue or keep us informed. But a few have taken our money, failed to provide a working service, then not responded for days – which is totally unacceptable and a perfect example of hosts that don't deserve your trust.

The other major problem area is a host which doesn't deliver on its promises, for example claiming a hosting plan includes feature X, when it really doesn't (or there's some other huge catch it's 'forgotten' to tell you). That's not just untrustworthy, it's dishonest, and earns the host a major black mark in any review.

Performance and reliability

Our web host speed tests are generally based on the cheapest shared hosting plan available from a provider. (In some cases, where we're reviewing a free hosting service, we'll use the free plan. If so, we'll spell that out in the article.)

We then upload a simple static site to our web space; just HTML and CSS files and some images.

Where possible, we point a test domain to our site. If this isn't supported, as is the case with some free plans, we use whatever subdomain the plan provides.

Next, we configure Uptime.com to check the availability and server response time of our site at five-minute intervals.

One week later, we note the site uptime (the percentage of checks where the server returned HTTP code 200, meaning 'OK') and average, minimum and maximum response time. We save the seven-day response time chart, too, as it's a useful indicator of consistency.

These results are handy as a way of spotting hosts with significant reliability or performance issues, but note that this testing also has significant limits.

We only test the cheapest shared hosting plans, for instance. The figures tell us nothing about the results you'll see with VPS or dedicated products, or any other plans which give you more resources.

The tests also take no account of page load time, and give us little or no indication of available CPU time, database setup, PHP configuration and more. We may partly address that within a review by using other tests, but there's no single figure that can usefully measure server speed.

Looking at the top and bottom-ranked hosts gives some indication of the best and worst performers, then, and points to web hosts most likely to deliver what customers need. But if you sign up for a provider, you should still run your own speed tests to confirm the service works for you.

Author

James Capell