Web Hosting Server Performance Tests: How To Do It Right

You might think that feature lists, performance speeds, and other web hosting review information are largely objective and reliable because they’re based on tangible data. However, that’s not the case. Many reviewers mislead readers to make their own jobs easier.

Reviewing web hosts isn’t easy. You need to pay for the service, run performance tests, and analyze the services so you understand who it’s best for. It’s a lot of time and effort, which is why some reviewers work with the web hosts directly. You’ll also find issues like falsified data or even rigged tests that promise amazing performance.

At Website Planet, we understand that your time and money are at stake. That’s why we take steps to ensure all our reviews are trustworthy. For one thing, we never divulge our identities to the web hosting services we test. We also focus on baseline performance tests to ensure we’re providing data that will be relevant to the average user.

It’s important to dig into the unpleasant details about reviewer-host collusion. This way, you’ll know how to spot an untrustworthy review, as well as what makes Website Planet’s testing process different.

The Problem With the Reviewer-Hosting Relationship

Review websites are, for the most part, businesses that need to make money. One of the most common ways websites make money is from affiliate commissions. Reviewers place an affiliate link in their articles, you click the link to make a purchase, and the website receives a commission.

When the affiliate system works the way it should, reviewers like me seek out and recommend great services that you’ll want to use. That’s a win-win for everyone: the host, the reviewer, and you, the reader.

The problem with the system is that some reviewers will take any affiliate deal that pays well. From there, they push their affiliates to customers without any regard to quality. In some cases, the web hosts will collude with reviewers. This can include feeding them data and talking points to use in their reviews.

The host may also provide special treatment, such as prioritizing shared resources for a review testing site. If a reviewer can’t preserve their anonymity from the hosting service, they can’t conduct reliable tests.

As previously mentioned, some review sites are more subtle with their dishonest coverage, as they’ll rig performance results to entice you into purchasing a subpar service. While technical data is objective, reviewers control the tests, so there are many ways they can manipulate and misrepresent the data they get.

How Data Can Lie (And Why Only Regional Results Count)

Assuming you’re conducting legitimate performance tests, you can’t directly increase or decrease a hosting service’s speed. While this basic speed is objective, it doesn’t represent real performance. There are several factors at play:

• When you attempt to access a web page, the page will load from the data center hosting it. The further away you are from the data center, the longer it takes to load.
• Once you build your website, it’ll have interactive elements, images, and other content that slow down loading times.
• A well optimized website will load more quickly than a poorly designed website.

It’s important to note that for most websites, the target audience is probably wider than the local area around your data center. Users will never load your site as quickly as a neighboring server will load an empty web page, but that’s how many reviewers conduct their tests.

By manipulating location and web page size, scammers can make the worst web host seem like it offers blazing speeds. Hosts don’t have a single speed that applies everywhere – real-life performance varies from country to country.

Here at Website Planet, all our testing is explicitly country specific. For instance, Hostinger is our top host1, but we don’t hesitate to say that other services are better in certain regions. We’ve tested it in multiple countries and found that Hostinger isn’t always the fastest.

For example, when we ran tests in the Netherlands, we found that local web hosts outperformed Hostinger in and around the country. Not only did we do the work to discover this, but we also made sure our readers knew about the alternatives.

This transparency and attention to detail is one of several pillars in our approach to web hosting. We account for distance, web page size, and other variables while using several tools to conduct realistic, dependable tests. Here’s how we do it.

The Right Way to Conduct Web Hosting Server Performance Tests

First of all, our reviewers never divulge their identity to the hosts. We operate as anonymous, everyday customers who get the same standard service you’ll receive. This way, we preserve our credibility and reliability as a source of hosting information.

Then, we purchase a hosting plan. To help give you an idea of a service’s baseline performance, we usually test the cheapest shared hosting plan available. If we choose a more advanced plan, we’ll say so explicitly in the review.

From there, we upload our standard testing page with images and interactive elements to the server. We run our testing servers from locations that are in the local country or region, but not too close to the host’s data centers.

All our testing runs “out-of-the-box,” unless we specify otherwise. That means that we don’t perform any complex optimization you might not be able to do.

We also don’t perform “stress” tests, which give you an idea of your website’s performance and stability under pressure. Unfortunately, it’s simply not possible to perform a stress test – such as attempting to overload a server – without contacting a host first to ask if it can remove certain security measures. Since doing so would reveal our identities to the host, the data we collect would be meaningless.

These steps are why our testing is a realistic, trustworthy measure of the baseline performance you can expect from a host.

UptimeRobot

Uptime is the percentage of time that your website is available to visitors online. Respectable web hosts typically guarantee 99.9% uptime or better, with refund offers for excessive downtime.

We use UptimeRobot to send our test websites HTTP requests every minute. This provides an extremely accurate picture of how stable and consistent a host’s services are. While some hosts fall short of their uptime guarantees, others come out above and beyond with 99.99%+ scores.

GTmetrix

GTmetrix has a worldwide network of test servers we use to automatically load our sites multiple times a day. GTmetrix records the results and provides detailed breakdowns with statistics covering time to connect, time to load the largest visual elements, and more.

Of course, not all GTmetrix reports are the same. Performance varies with how busy the server is, so results can change from one test to the next. That’s why we run these tests for at least a week to ensure they’re accurate.

By the end of our GTmetrix testing, we have dozens of reports from which to draw insights. Each of our reviews with testing include graphs of performance over time, as well as a sample report that represents a normal performance level for the host.

The Sucuri Load Time Tester

Sucuri Load Time Tester utilizes a global network of testing servers to analyze a website’s performance worldwide. The tool logs the time it takes to make the first connection and the time it takes for the browser to receive the first byte of data from the server, as well as the total loading time.

It tests performance separately for each location and offers an overall global performance grade. With Sucuri, we’re able to understand how well a host performs around the world and how well its built-in caching optimizations work.

Best by Test: Performance Results You Can Trust

We’ve written numerous articles comparing the best web hosting options for different budgets and needs. All our reviews on performance rely on our rigorous, consistent testing methodology that gives you real results. Our integrity and attention to detail translate into reviews that you can trust.

In addition to the performance criteria above, we also test each hosting service across several other categories:

• Features. This includes the resources, tools, services, perks, and other bonuses provided with each hosting plan.
• Ease of Use. A more subjective measure of a host’s overall user experience, we look at everything from the sign-up process to the user interface.
• Support. When questions or concerns inevitably arise, it’s important that your host can provide you with the answers you need. We test each host’s various support channels so that you know what kind of customer service you can expect.
• Pricing. The main concern we cover here is how much value you get for your money. But we also look at available payment options and money-back guarantees.