Early findings: Internet Explorer 9 will not make your web site faster by default

As with any new browser release, there’s a lot of hype around Internet Explorer 9. And there’s particular hype around IE9’s ability to deliver a faster web experience. Here’s what Microsoft has to say about it:

“Hardware-accelerated text, video, and graphics mean your websites perform like the programs that are installed on your computer. High-definition videos are smooth, graphics are clear and responsive, colors are true, and websites are interactive like never before. With sub-system enhancements like Chakra, the new JavaScript engine, websites and applications load faster and are more responsive.”

I’ve been wondering about the impact of IE9 on our business, so I decided to undertake a research project and try to quantify how much better it is than previous IE versions in terms of performance.

Methodology

  • Take the landing pages of the top 200 Alexa-ranked retail sites
  • Run tests in IE6, IE7, IE8 and IE9 (including video) on Webpagetest from Dulles, VA
  • Record and compare the results

Initial conclusions

As you can see on these graphs, IE9 has very little impact – and possibly even a negative impact – compared to IE8, in terms of load time and time to start render.

Internet Explorer: Average first view load times for the Alexa Retail 200Internet Explorer: Average first start render times for the Alexa Retail 200

Further conclusions

When I shopped my findings around Strangeloop, I got this response from our CTO, Kent Alstad:

“My conclusion is that IE9 is mostly aimed at HTML5 (especially CSS drawing) and JavaScript acceleration. I suspect that these sites [the Alexa Retail 200], in attempts to be faster and more usable to more people, do not exploit these enhancements. I think some comparison with sites that use the new features would show a more dramatic improvement.”

IE 9 is not the problem. It’s actually a good browser. I like it. But it won’t make websites faster by default — and to its credit, it doesn’t claim to do this. This is merely a conclusion that the general public has leapt to.

It’s crucial to bear in mind that a browser is just an application that renders a site to visitors. To take advantage of IE9’s potential to deliver faster performance, it’s up to site owners to optimize their sites so that IE9 can render it as quickly as possible.

IE9 will offer a dramatic improvement when more sites move to HTML5 and Ajax, but right now, when fewer sites are using these techniques, the gains are not present.

Related posts:

17 thoughts on “Early findings: Internet Explorer 9 will not make your web site faster by default

  1. Pingback: Tweets that mention Early findings: Internet Explorer 9 will not make your site faster by default — Web Performance Today -- Topsy.com

  2. Can you explain the significance of using “start render” and “load time” to measure the stuff mentioned in the first quote?

    Hardware acceleration is talking about local hardware – it can only act on media after it comes across the wire. It’s smoother playback, not increased download speed.

    And modern JavaScript doesn’t run (sometimes doesn’t even download) until after page load time. Do you have graphs showing how fast Trident parses/executes big scripts versus how fast Chakra does it?

    I agree with your conclusion, I just feel like I’m missing the steps that got there and don’t have any data myself to connect the dots.

    Thanks!

  3. Hi Josh!

    In your comparison with IE 9 do not forget, that for whatever reason, according to browserscope.org, the Browser opens up only 2 connections per hostname. In contrast to IE 8, which does 6.
    It would be interesting to know, if they will switch to 6 connections with the final version of IE 9. Or if they don’t, why.
    Josh, if you have some spare time, maybe you can re-do the test with IE 9 and 6 connections per hostname? :-)

    Kind regards,
    Markus

  4. Erratum:
    I just saw, that with IE 9 BETA, they switched to 6. And as this is the IE 9 version, that webpagetest.org has installed, your tests were indeed done with IE 9 doing 6 connections per host name. So simply ignore my last comment.

    Kind regards,
    Markus

  5. Fascinating, thanks. Not really in the least bit surprised.

    We have similar misgivings about the feature-set they’ve tackled. There’s lots of hardware accelleration but they’ve missed out lots of [boring] things that would make general web development more effective. There’s a bit of a write-up on the feature set here: http://goo.gl/xgQG6

    J.

  6. Pingback: IE9 Lags on Popular Websites: Report | Find Tech News

  7. Pingback: IE9 Lagi na popularnych stronach internetowych: Sprawozdanie | plane

  8. Pingback: Speed: Recommended for Content, Conversions and Holiday Traffic | Services For Seo

  9. Pingback: OPISO » Speed: Recommended for Content, Conversions and Holiday Traffic

  10. Did you find that on the first page load after a DNS flush that the main request was being block by all sorts of DNS calls happening before the main sites get request?

  11. Pingback: Speed: Recommended for Content, Conversions and Holiday Traffic | Programming Blog

  12. Pingback: Speed: Recommended for Content, Conversions and Holiday Traffic | O-I Newswire

  13. Pingback: World Spinner

  14. Pingback: End-of-year performance report: Top retail sites are slower, not faster, than the rest of the pack — Web Performance Today

  15. Pingback: IE9 Lagi na popularnych stronach internetowych: Sprawozdanie | inverse

  16. Pingback: Performance Calendar » The State of Web Performance Optimization – 2010

  17. Pingback: Is it time to give up Internet Explorer 7 as our default test scenario? — Web Performance Today

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>