How?

Instead of relying of a small sample of reviewers who might not have your best interests at heart, at hosting scorecard.com we rely everyone who has ever expressed an opinion of the hosting company in question.

The internet is full of people with opinions.  Every day tens of thousands of people with no financial or other incentive will publish their opinion. They do it on blogs, they do it on forums they do it on mailing lists and newsgroups.  Sometimes they do it in response to a question; sometimes they do it because they are very happy or unhappy with their current host.  Regardless of where or why these are real people with real opinions.  These are the people that you want to ask, all of them, not just a few.

That’s what we do at hosting scorecard.  We look at everyone opinion not just a few.

How does it work?

Have you ever been to the website sucks-rocks.com or one of its variations?  The idea is that you query a search engine for the phrases "SomeCompany sucks" and "SomeCompany rocks" and use the number of results for each as an indicator of the online reputation of SomeCompany.

We have taken that idea and expanded on it, we look at 18 different keywords and how they are used in association with different hosting companies.  We also break the searches down into 1 week intervals so we can see the trend from week to week.

The big question has to be “Does it work?”

Let the data speak for itself.

Below is one of our hosting graphs.  This graph is a summary of 8900 different opinions that have been expressed over a 6 month period. This graph is specific to one hosting company, the searches used to create it contained the full domain name of that host. Each blue cross is the score for a particular week and the thick blue line is the trend for the last 6 months. The thin blue lines represent how consistent the weekly scores and are drawn 1 standard deviation on ether side of the main trend line. The red arrow is the final score for this host, in this case about 94. The vertical bars at the bottom represent volume.

One thing you would notice is the weekly scores are fairly consistent, most of them occur close to the dark blue trend line.  


Now let’s look at another graph.  This graph is a summary of 5200 opinions of a different hosting company that have been expressed over the same 6 month period.  The individual scores are again in a fairly narrow range round the trend line, but the overall score is much different, its about 66.



There is only one difference in the way data was collected for both these graphs.  The searches used to create the data for the graph each contains the domain name of a different hosting company.  The data was collected at the same time by the same methods.  Because the only difference in the data collection is the host name, then any changes in the data is because of the host name and not some other influence.   

Consistency is also important; the fact that there is not a wide swing in the weekly results would indicate that the results represent the real reputation of the host and not just a random fluctuation within the data.

We can therefore conclude that the first hosting company has a much better online reputation than the second one does.

Conditions for inclusion in the list.

In order to be objective it is also necessary to disclose the criterion used to determined which hosts are included on the list, its easy to exclude hosts to create a list that is just as biased as one that ranks hosts by affiliate payout.

When we started this project the first task was to create a large list of hosts.  We ended up with 10,000+ hosts.  We collected preliminary data on all of them.  We pared the list down to about 1200 based solely on volume.  An amazing number of those 10,000 had virtually no online presence and could easily be eliminated.  For remaining 1200 we collect data weekly.  The final list that comprises this site is extracted from those 1200 based on the following rules:

1. Must be in English

2. Must be a Host.

3. Volume.

4. Score.

5. Consistency.

These 5 rules are somewhat arbitrary, however they are consistent. Every host that passes those conditions will be included on the list and no hosts that fail any of the tests are included.  From time to time we might tweak the threshold values but we will always apply the rules consistently.

Note 1.

Note 2.

Note 3.