Problems with market share statistics.

Some time ago while researching for a project I needed to have an understanding of the major browsers and their market share. While browsing I came across 3 different sources. These were :

Upon checking the results I found that the results for each browser were wildly different between each site for example for the month of December 2013 are as follows:

w3schools

Internet Explorer – 14.7%
Firefox – 31.9%
Chrome – 46.9%
Safari – 4.2%
Opera – 2.1%

Stat Owl (stats are for November for some reason there were no up dates since then)

Internet Explorer – 43.38%
Firefox – 19.42%
Chrome – 24.91%
Safari – 9.27%
Opera – 0.75%

Market Share

Internet Explorer – 54.77
Firefox – 19.82%
Chrome – 18.04%
Safari – 5.24%
Opera – 1.71%

 

As you can see there are similarities between Stat Owl and Market Share however the w3schools stats are completely askew.

Wo what is happening? Are w3schools trying to deceive us? Well no, the issue is more mundane than that. From reading the page a little more carefully you will notice a paragraph that helps answer that question:

From the statistics below (collected from W3Schools’ log-files over a period of ten years), you can read the long term trends of browser usage.

Now let us consider the type of site that w3schools actually is and who it is targeted at.The type of content that w3schools has is clearly aimed at people who create websites and lets face it people who create websites tend to be more along the tech savvy type of user. From experience I know that tech savvy people are not your typical web user. Tech savvy users are more likely to choose the browser they use and not just use the 1 that comes as standard.

Considering this the w3schools are all but useless unless you are interested in the current habits of tech savvy individuals rather than the general populace.

Now we move onto Stat Owl. The first issue that we come across is that the data is somewhat incomplete. The tables (at the time of writing) show that the only browser that has any data since November 2012 is other therefore they have no valid data for the last 6 months.

The data that Stat Owl does have however does at first glance look to be quite believable. Now we saw the issue with w3schools so let us have a look as to where Stat Owl obtain their information from.

Helpfully Stat Owl have a page specifically giving us the information we are looking for on their How We Collect Data page. On a positive note Stat Owl claim that they try to remove as much noise traffic as they can (such as bots and search engines). The data they collect is also based upon an average of 28,000,000 views a month on a network of sites. There is however a slight downside to the information that Stat Owl provide. Stat Owl give a quick summary of the data that they have and advise that 80% of the sites they receive data for a predominantly US centric sites. There is nothing inherently wrong with this but can essentially be biased and not reflective to the world as a whole. Stat Owl also do not log mobile interactions which is becoming more and more important, the fact they do not have this at this stage of the game is quite worrisome. We are well past the stage of the mobile market exploding. The big bang happened some time ago.

Our last provider is Market Share. I am quite impressed with Market Share. They have a wealth of reports and break down the results for both mobile and desktop. As with Stat Owl the data they have appears to be believable. There are however some downsides. Although it appears that there is a vast array of data points Market Share does not appear to give any information about where the data is received therefore we do not know if this data is a good representation (there is a ticker banner at the bottom of some pages that shows logos of some very big name sites but no explanation why, I can only guess the data comes from these sites). Quite a few of the reports are actually chargeable as well.

SUMMARY

It is needless to say that if you are interested in understanding which browsers are mainly being used and which browsers to target, take any information you get with a pinch of salt. You should also look into where this data comes from and how reliable the data appears to be.

If I were to choose 1 of the services mentioned here I would have chosen Market Share but I would be a bit concerned about where the data has been retrieved from.

There is of course 1 alternative. If you run your own site and have a decent amount of hits you could always use tools such as Google Analytics. This will of course give you information about the traffic that means the most to you, your own customers.

Tell us your thoughts