Our test methodology

The essentials first

Our test reports are meant to give you a transparent, practical basis for choosing the right web hosting package or VPS. Three principles apply:

• Tested anonymously

  The hosttest editorial team carries out every test anonymously. That way we experience ordering, setup and support exactly as you would as a customer.

• At our own expense

  We order every product ourselves and pay the normal customer price, just like any other customer.

• Independent and neutral

  Providers have no influence on the process or the result. There are no paid test results and no financial influence from outside.

How a test report is made

So that offers from different providers can be compared fairly, every test follows the same standardised process:

1. Choose a provider

   We pick a provider that customers are interested in, from well-known big brands to smaller, owner-managed hosts.

2. Narrow down the product

   We look at the provider's range and choose the offer we will test. To keep the comparison meaningful, we look for a similar feature set and price range as in our other tests.

3. Pick the plan

   We compare the features of all plans and check the key contract details such as term and notice period.

   Web hosting

   Packages at roughly €2 to €5 a month that can run WordPress.

   VPS

   Offers at roughly €5 a month with 2 vCores, 2 GB RAM and SSD storage.

4. Order, set up, measure

   We order the product, set it up and document every step from ordering to the finished system. Then we measure its performance.

   Web hosting

   We create a WordPress website and measure availability and loading times.

   VPS

   We install Ubuntu and measure CPU performance and disk speed.

5. Rate and publish

   Based on our hands-on experience we write the test report and rate the product in five categories.

The five rating criteria

Every test report follows the same fixed rating criteria, so you can compare the offers directly. We award up to 5 points in each category.

• Features

  max. 5 points

  We look at the scope of services and further features in the areas of technology and security.

  Web hosting

  Storage, databases, included domains, external domains, free SSL certificates, custom PHP settings

  VPS

  vCores, RAM, storage, choice of operating systems, backups, DDoS protection, firewall settings

• Usability

  max. 5 points

  We walk through the order process step by step and check how easy the control panel is to use. Payment options and the time until the service is ready also count.

  Web hosting

  We also document and rate installing and setting up WordPress.

  VPS

  We also rate how easy it is to manage the VPS in the customer panel.

• Performance

  max. 5 points

  We measure performance with fixed tools at three points in time within a month. Which metrics we use and how they turn into points is explained in the section on performance measurement.

  Web hosting

  Availability, response time and loading time, with a focus on fast access from Germany

  VPS

  CPU performance (single and multi core) plus disk speed and IOPS

• Customer service

  max. 5 points

  We check the provider's help and FAQ pages, the direct ways to get in touch and our own experience with support.

• Costs

  max. 5 points

  We rate value for money: choice of plans, optional extra costs and discounts such as limited-time offers or free trials.

  Web hosting

  We also take into account possible savings from included domains and figures such as storage per euro.

From points to the overall rating

For the verdict we add up the points of the five categories and divide the total by five. This gives an overall rating from 0.0 to 5.0 points, which we show on a 5-point scale.

1. 5 categories0–5 points each
2. Totalmax. 25 points
3. ÷ 50.0–5.0 points
4. Overall ratingout of 5

Since December 2025, customer reviews no longer count towards the overall rating. You will still find them in the test report and the provider profile.

How we measure performance

So that a single snapshot does not distort the result, we measure every value at three different points in time within a month and take the average. Each metric is classed as good, medium or poor.

Web hosting

• Uptime

  The share of time in which server and website are reachable without outages or interruptions.

  Measured with: our own monitoring tool

  • good99.5% or more
  • medium98–99.5%
  • poorbelow 98%

• TTFB (Time to First Byte)

  The web server's response time: the time between the request and the first byte of the response.

  Measured with: Uptrends Website Speed Test (desktop, Chrome, Frankfurt)

  • goodbelow 300 ms
  • medium300–500 ms
  • poorabove 500 ms

• LCP (Largest Contentful Paint)

  The time until the largest image or text block in the visible area is rendered, usually the main content. Google simulates real users on a throttled connection, so the values are higher than in other tests.

  Measured with: Google PageSpeed Insights (Europe, throttled mobile connection)

  • goodbelow 1,500 ms
  • medium1,500–2,500 ms
  • poorabove 2,500 ms

• Page speed

  The loading time: how long it takes until the whole website has loaded in the browser.

  Measured with: Pingdom Website Speed Test (Europe, Frankfurt)

  • goodbelow 500 ms
  • medium500–1,000 ms
  • poorabove 1,000 ms

VPS

• Single-core CPU

  The performance of a single CPU core. Important for applications that mainly run on one core.

  Measured with: Geekbench (in the Yet-Another-Bench-Script)

  • goodabove 1,250
  • medium1,000–1,250
  • poorbelow 1,000

• Multi-core CPU

  The performance of the CPU when several cores work at the same time. Important for parallelised applications.

  Measured with: Geekbench (in the Yet-Another-Bench-Script)

  • goodabove 2,000
  • medium1,500–2,000
  • poorbelow 1,500

• Disk: read/write

  How fast data can be written to and read from the drive.

  Measured with: Flexible I/O Tester (fio)

  • goodabove 1,500 MB/s
  • medium1,000–1,500 MB/s
  • poorbelow 1,000 MB/s

• Disk: IOPS

  How many input/output operations the disk handles per second. Important for many small accesses, for example from databases.

  Measured with: Flexible I/O Tester (fio)

  • goodabove 25,000 IOPS
  • medium15,000–25,000 IOPS
  • poorbelow 15,000 IOPS

From measurements to stars

Each of the four metrics earns points: good = 3, medium = 2, poor = 1. Together that makes 4 to 12 points, which we convert into a star rating.

Our code of conduct

• Every test report shows both the strengths and the weaknesses of an offer.
• We regularly check the tested offers for changes and make updates transparent.
• Our test results are not for sale. We refuse payments from providers that are meant to influence our rating.
• We hold no stake in any tested company and have no external investors, so no conflicts of interest arise.
• Feedback is welcome: customers and providers can comment on our content.

All test reports

We publish new test reports every year. These providers already have a detailed test, sorted by overall rating:

Web hosting (5)

• DreamHostTested: 02/2026 4.8/5Excellent
• IP-Projects GmbH & Co. KGTested: 02/2026 4.8/5Excellent
• ALL-INKL.COM - Neue Medien MünnichTested: 04/2026 4.6/5Excellent
• Cloud86Tested: 02/2026 4.5/5Excellent
• netcupTested: 02/2026 4.2/5Very Good

Go to the Web hosting comparison

VPS (7)

• FasthostsTested: 01/2026 4.8/5Excellent
• netcupTested: 01/2026 4.8/5Excellent
• Contabo GmbHTested: 01/2026 4.8/5Excellent
• HostingerTested: 09/2026 4.6/5Excellent
• IP-Projects GmbH & Co. KGTested: 01/2026 4.6/5Excellent
• Hetzner Online GmbHTested: 01/2026 4.6/5Excellent
• InfomaniakTested: 01/2026 4.2/5Very Good

Go to the VPS comparison

Keeping test reports up to date

The hosting market keeps changing, and providers adjust their packages all the time. That is why we usually update our test reports once a year in January: we check features, terms and the other test criteria again and update the report when something has changed.

Changes to the test methodology

1. 3 Nov 2026Test methodology for web hosting and VPS merged into one page and measurement thresholds clarified.
2. 16 Dec 2025Test methodology revised: customer reviews no longer count towards the overall rating.
3. 7 May 2024Information on the test procedure and the performance test updated.
4. 1 Sept 2021Test methodology first published.