Tag Archives: browser wars

Good news on IE6: government usage plummets in 2012

Almost a year ago, we published some interesting browser statistics based on the logs from one of our Central Government Citizen Space servers. We ran the logs through a piece of open source analysis software called Visitors, and this gave us an anonymous breakdown of all visits to Citizen Space, showing the browsers and versions that were used. We looked at the statistics for all pages, and compared them with the stats for pages only accessible to admin users. The results were insightful but rather scary: Internet Explorer 6 accounted for more than 1 in 3 visits by our Central Government admin users.

For those who don’t know, Internet Explorer version 6 (lovingly known as IE6) is a web browser that Microsoft released over a decade ago. Because it renders web pages differently (in some cases dramatically differently) from more modern browsers, web developers spend a great deal of time creating workarounds so that IE6 users can still access our websites. Of course, this increases the cost of product development without necessarily offering any benefit to the majority of web users who don’t use IE6. Coupled with the fact that IE6 now only receives limited support from Microsoft, almost everyone is in agreement that this ancient and decrepit browser must be phased out.

This morning, we had a comment on last year’s post from a reader who was interested in how Citizen Space’s browser stats had changed. Thank you Perry – you reminded me that I’d been meaning to re-do this analysis soon. So here are the graphs comparing the numbers 11 months ago with where we are today:

Central Government Citizen Space – all users (admin and public)

All users by browser, May 2011
Breakdown of visits by browser: IE8: 23.0%, Firefox: 15.6%, IE9: 14.9%, IE7: 14.9%, Chrome: 13.2%, Safari: 10.8%, IE6: 5.9%
All users by browser, April 2012

These statistics roughly follow the browser trends of the general internet population*, with IE8 and 9 increasing in popularity while the older IE versions decrease as expected. Pleasingly, IE6 usage has roughly halved since last May.

Firefox, Chrome and Safari have gained more of a stronghold in the past year, although interestingly, Internet Explorer as a whole has retained a far larger share of Citizen Space users than worldwide browser usage statistics* would predict.

Central Government Citizen Space instances – admin pages only

Admin users by browser, May 2011
Breakdown of visits by browser: IE8: 60.3%, IE7: 29.3%, Chrome: 3.8%, IE6: 2.4%, Firefox: 1.8%, IE9: 1.3%, Safari: 1.1%
Admin users by browser, April 2012


When looking at the statistics for our admin users, the most exciting thing is that usage of IE6 has crashed by 90% – from 35% down to 2.4% of visits. This is a great relief to us, and shows the huge effort that has taken place in government IT departments to upgrade users away from this insecure, ill-supported browser.

It’s worth noting that overall, usage of Internet Explorer among our Central Government users is more than 90%, compared to 34% worldwide*.

What next?

The interesting question is what levels of support to provide for different browser capabilities. We currently provide Level 2 support for IE6, which means that all content must be readable and navigable, but differences in styling and layout may exist. This works OK for our products at the moment, but as web users come to expect a richer and more fluid experience, the likes of IE6 are going to lag further and further behind. How small does the percentage of IE6 users need to be before we can stop supporting it at all?

To other web developers: when do you stop supporting ancient browsers and those with limited functionality? How much can you rely on the presence of client-side technologies like Javascript, cookies, HTML5, CSS3..?

To government IT managers (thank you for phasing out IE6 btw!) what level of support do you expect for older browsers? Do security constraints dictate that you disable features like Javascript or cookies?

As always, I’d love to hear your views.


*Worldwide browser statistics from statcounter.com.

IE6 – the facts: 1 in 3 of our government clients still using it

We spend a lot of development time ensuring that our software is usable across all the commonly used web browsers. A disproportionate amount of this time is spent ensuring compliance with Microsoft’s archaic Internet Explorer 6 (IE6), which is 10 years old this year.

From time to time, we wonder whether we could provide better value for money by dropping support for IE6 and making more use of the facilities provided by newer browsers; a decision that has already been taken by the likes of YouTube, Facebook and Google.

Steph Gray recently blogged about whether Alphagov should have dropped IE6 support. Steph critiqued this decision, pointing out that a lot of civil servants still use IE6. We thought it might be useful to share the breakdown of browser usage by our civil service clients.

What do the numbers show?

Here is a chart of browser usage on the admin pages of our Citizen Space app, for the servers used by our Central Government clients:

IE7: 47.5%, IE6: 35.4%, Firefox: 7.1%, Chrome: 5.9%, IE8: 3.4%, Safari: 0.7%
Breakdown of admin pages by browser

And here is a graph of browser usage across all Citizen Space pages (admin and public-facing) on the same servers:

IE7: 21.2%, IE8: 20.3%, Firefox: 14.3%, Unknown: 13.2%, IE6: 11.0%, Chrome: 9.3%, Safari: 6.6%, IE9: 3.0%, Others: 1.2%
Breakdown of all pages by browser

Here you can see that IE6 is used by more than a third of our Citizen Space administrators, but only about a tenth of the total visitors. At the moment, there is clearly a need to continue supporting IE6 for our clients, but it does seem a shame when this investment could be put towards improving the user experience of the site’s end users.

I could write more words about these differences, but here’s another chart that tells the story pretty clearly:

Comparison of browser usage by admin users and all visitors

We’ll be keeping an eye on these figures to see how they change with time, and we’d also be interested to know how they compare with data from other sites aimed at government clients around the world. Does anyone else have any data they’d like to share?


The science bit:

  • Data was taken from May 2011’s Apache access logs from one of our Citizen Space servers. Data is anonymised.
  • We parsed the logs using the open-source Visitors software (which we modified to include the most recent versions of IE). The software can be downloaded from http://www.hping.org/visitors/doc.html.
  • Statistics are based on visits rather than pageviews, where a visit is all requests for a given useragent and IP address on one day.
  • We excluded any visits from our own IP address and from our server monitoring services.
  • We did not exclude crawlers and other bots, which probably account for the majority of the ‘unknown’ useragents in the second chart.