What is HRank?

HRank is the web hosting rating system based on research, analysis and experience and is meant to be objective and up-to-date; and free.

HRank analyzes only shared web hosting providers.

more than 300 web hosting providers
16,000 shared IPs*
– 10 mln websites**

*1 Shared IP contains more than 150 websites

** We count number of websites with “150 and more domains rule”. It means that we take only Shared IPs with >150 websites – it allows us to cover the maximum number of physical servers and domains they host; and to provide the most reliable uptime information.
We have data about ~150mln websites. We check which web hosting provider they belong to on a daily basis.
However, there are only ~10mln websites that comply with our “150 and more domains rule” (except for Parked Domains). And all Shared IPs (Servers) with these ~10 mln websites we ping daily for their uptime and response time.


As opposed to the majority of existing hosting review companies that base their decisions on the uptime check option, we do not offer any software or online tool to monitor your personal website. We are for a bigger and much more objective picture when it comes to shared hosting performance. Our rating system estimates a hosting company on the basis of several factors, the major of which being

We provide the most accurate and reliable data with regards to hostings’ past and current performance – how they worked for a period of time, how long and how often they were down and how fast their Shared IPs respond.

The research and analysis we’ve done since June 2018 are really huge and it is in our plans to do so much more.


How We Do the Monitoring

  • We monitor and collect data from the websites on Shared IPs (constantly).

Shared hostings can be (rarely) sparsely populated; however, most of the time they are pretty crowded (some may host up to a thousand websites.) After all, that’s the whole purpose of a shared hosting service – cheap and cheerful.

  • Our software aggregates and analyses the data and provides uptime and response time information for the domains on every Shared IP (the average value for the previous 30 days).

We know for sure what’s going on with all TLDs on the Internet (in the most popular zones e.g. .com, .net, .org and etc.). We check their IPs and nameservers, and we also know which provider they belong to. The information about the providers and the number of their domains is updated EVERY DAY.

Then we make a logical step is to correlate the domains – their IP and provider nameservers. We group the identical ones using our “>150 domains” rule, and this is how we come to the number of shared IPs that you can see on a provider’s page.

In our work we have had to apply “>150 domains” rule – we take only Shared IPs with >150 domains. Such an approach allows us to cover ~99% of the physical servers and domains they host. By doing so we are able to provide the most reliable uptime charts you can find anywhere on the Internet at the moment.

We have data about ~150mln websites. We check which web hosting provider they belong to on a daily basis. However, there are only ~10mln websites that comply with our “>150 domains” rule. And all Shared IPs (Servers) with these ~10 mln websites we ping daily for their uptime and response time.

Our system sends 288 “get http” attempts (pings) to the domains on each Shared IP every 24 hours.
Every attempt can consist of up to 5 requests to understand whether a server/Shared IP is available or not.

288 pings for siteground servers

Each interval between attempts is about 5 minutes for every Shared IP. The attempts are totally random, and they choose a new domain for checking every time.

example of domain for siteground server

If one domain on the Shared IP is unavailable, our system sends the request to the next one, and so on. Each next domain is always selected at random.

bad server request

After several unsuccessful attempts of our robot to access domains hosted on it (provided that the sites were active, say, half an hour ago), the decision about general unavailability of domains on a Shared IP is made and is further on reflected on a provider’s chart with downtime.

Here is how we check the domains on each Shared IP: 

  • we send a “get http” request to the 1st domain that worked (was active in our history) 5 minutes ago,
  • if it fails to respond, we go to the 2d site that worked (was active in our history) 5 minutes ago and send a “get http” request,
  • if it does not respond, we go to the 3d domain that worked (was active in our history) 5 minutes ago and send a “get http” request,
  • and so we do till we reach the 5th site.
  • if all 5 sites are inactive and fail to respond = > the system marks that the domains on the Shared IP have been down for 5 minutes and waits for the next round.

If may look like such an approach defeats the purpose, but it’s not true. This is done specifically for the sake of checking the domains on the Shared IPs that have recently stopped responding. Thus, you can always see uptime and response time statistics for the previous week – which is the most reliable data you can get. And we intentionally give the domains that seized to respond an extra week; if they don’t grab the chance, their Shared IPs are no longer taken into the account.

With all this, we are able to outline hosting performance characteristics much more precisely.


Uptime

  1. We calculate uptime on the basis of the data that we have at our disposal (the number of domains and Shared IPs).
  2. To calculate the HRank rating we take the average monthly value (the average of 30 uptime values, one for each day.)
  3. Data charts illustrate daily uptime value that is calculated every 24 hours in the following way:
    • Shared IP’s inaccessibility time is taken into account (all the domains hosted on this IP are checked for response), the share of each IP is calculated (with regard to the number of websites) and its % of inaccessibility in terms of a time period is identified. This allows us to draw the uptime % for a certain given day.

Response Time

The calculation of response time follows pretty much the same steps as described in the uptime section.

For a web hosting provider to be called “good”, its response time should be fairly short. A good average comes within 200-350ms range, and we check it exactly for every Shared IP (host) – taking into consideration all the factors that can affect or impair the number. Then, again, we draw an average from all the Shared IPs and it becomes another component of HRank.


Support

We have a considerable number of online projects, thus, we are sighed to more than 200 web hosting providers and we know about web hosting pros and cons first-hand.

We can provide an expert opinion as we do have, well, a tremendous share of experience and we also know how the companies work with their customers, in other words, how their support really works.

In cases when a hosting provider is unfamiliar to us, we initiate a phone call or a live chat conversation with tech support.

We represent ourselves as potential customers and make inquiries of both general and technical nature, ask tricky questions about technical aspects and not the overall support team approach (and speed).

So, eventually, we get a 3-dimensional picture that is both objective and up-to-date (the ratings are subject to change as soon as any alternations are detected in the process of monitoring). Our rating system gives an opportunity to look at the hosting performance history and draw results on the basis of multiple examples (instead of one, or several).

At the present moment we can provide you with the information about 240 providers and their shared plans, all of them we have experience dealing with.


Other Factors

AGE: We pay attention to the age of the website and how long the company has been in the industry. Web hosting market is quite versatile and there is no direct correlation between age and quality, though, sometimes companies that have been there for years are the ones that you can rely on. And sometimes not.

DESIGN: While good design does not necessarily mean good quality, outdated one is able to draw away even least fastidious customers. Companies that make an effort to keep their website nice, up-to-date and user-friendly are definitely ranked higher in our rating.

THE REST: Finally, we go and check many other details. Better performance web hostings with best profiles are rated higher in HRANK.


We do not try to:

  • draw you to some particular hosting;
  • discriminate any hosting provider by presenting unfavorable reviews on purpose.

We try to:

  • be objective;
  • and of use.

We have no aspirations for being hosting gurus, however, there is a lot we can say on one particular subject. We specifically focus on shared web hosting as it is basically one of the most popular among website owners like us – not enterprises, but individuals and small businesses.