It may be time to stop supporting Internet Explorer 10.

That probably sounds a bit crazy. I’m writing as 2015 begins, which means that IE 10 was the very latest version of Internet Explorer available to Windows 7 users just 14 months ago. But the intervening year has made one heck of a difference.

Today, for every one Web visitor running IE 10, you will serve four and a half users of IE 11. 1

Amazing! There is a minor caveat, however. The somewhat counter-intuitive fact is, there are more IE 9 users out there than IE 10 users — slightly. You’ll see one and a quarter IE 9 users for each IE 10 user. And… there are even more IE 8 users, by an even smaller margin. It works out to less than one and an eighth IE 8 users for each IE 9 user, or almost one and a half for each IE 10 user. 2

There are still more IE 11 users than IE 8 + IE 9 + IE 10 users, combined; it’s a 6:5 ratio. In other words, you can reach 55% of our friends using brain-damaged browsers by targeting only the one latest version of IE.

Even in the US, where Internet Explorer is used much more often that the worldwide average, less than one out of every 37 visitors will be using IE 10. Would you rather use the latest browser technology to give 36 people a better web app, or one slow upgrader an "equally good" experience? Before you answer, ask yourself: Am I a communist? and, Do I believe in limiting myself to the common denominator? After ditching IE 10, features you can start using without any polyfills or fallbacks include the hidden attribute, modern flexbox with no fallbacks, CSS border-image, WebGL (3D canvas) (sorry, non-Chrome Android browser), PNG-only favicons, cryptographic-strength random values, the full screen API for desktop, MutationObserver, SPDY/HTTP 2 (mostly, and only on Win 8.1), and even the Internationalization API (if you don’t mind punishing Safari users for Apple’s decision to make the only current browser without it).

How to decide? The one key number here is 2.8%. That’s the maximum number of potential visitors you lose by bumping your minimum browser up from IE 10 to 11. It will be even less if some of these people solve the problem by whipping a smartphone out of their pocket, grabbing their tablet from by the bed, or launching Chrome or Firefox.

Of course, it comes down to trade-offs — IE 10 supports most of the things you want a browser to, so there’s a good chance your app won't take much work to become compatible. But if that isn’t the case? Unless you’re building a service that must work for every single citizen, don’t feel so bad about it! IE 10 will be virtually dead in a year.

Wherever your organization draws the line, it’s abundantly clear that IE 10 will not generally require support for anywhere near as long as IE 6 did. And that’s something worth celebrating. We avoided the doomsday scenario.

  1. Source: StatCounter; other sources of market share numbers agree with the ratios I reported in this article. ↩︎

  2. Note that the ratios of IE 8-vs-9-vs-10-vs-11 usage hold both within the United States and worldwide. (IE use is more prevalent within the US than elsewhere.) ↩︎

February 6th, 2015
Alan Hogan (@alanhogan_com).  Contact · About