Experimental new Webpagetest feature lets you test sites on Chrome
23 Mar 2011
I’m always excited any time WebPagetest announces a new feature, and the recent announcement of the experimental new Chrome feature is no exception. At the risk of being labelled “that guy who runs a bunch of tests every time a new performance measurement tool comes out,” I… well… I ran a bunch of tests.
The test subjects were some of the winners of Gomez’s “Best of the Web” performance awards. I wanted to see how they performed across all of WPT’s current browsers, from the Dulles, VA, location. Here are the results of a simple load time test:
| Website | IE7 | IE8 | IE9 | Chrome |
| BB&T | 2.182 | 2.688 | 1.653 | 2.279 |
| Regions Bank | 1.876 | 2.489 | 1.736 | 1.515 |
| Capital One | 5.941 | 4.556 | 6.075 | 5.017 |
| Fidelity | 1.882 | 2.651 | 1.853 | 1.621 |
| United Health | 4.032 | 2.001 | 1.995 | 1.917 |
| CBS News | 15.354 | 15.335 | 12.668 | 15.047 |
| Newegg | 7.757 | 4.543 | 4.684 | 4.374 |
| QVC | 4.012 | 3.275 | 3.340 | 5.276 |
| Delta | 5.017 | 3.846 | 5.302 | 4.406 |
| Radisson | 9.779 | 11.530 | 8.016 | 8.560 |
| JetBlue | 9.654 | 7.767 | 5.580 | 6.083 |
| FEMA | 2.922 | 2.587 | 2.407 | 2.432 |
No big surprises or inexplicable anomalies here. For example, when I looked at the waterfalls for the sites that seemed to perform better in Internet Explorer 7 than they did in Chrome, the issues seemed to be code-related, not test-related.
Overall, the Chrome tests worked great, albeit in a currently limited capacity. (Right now, these are a few of the features that are not yet active: optimization checks, HTTPS, SPDY, content blocking, and above the fold time.) As Patrick Meenan notes in his announcement on the WPT forums, the new Chrome feature is “still very much in development, but I think it’s far enough along to be valuable and to start collecting feedback from you guys.”
All in all, this is a great new development — the first milestone in the three-stage roadmap that Pat announced in January. I’m looking forward to the launch of the Firefox and Android tests.
Mar 23, 2011 @ 15:58:29
Just a quick heads-up, if you ran your tests before this morning there’s a chance that the chrome times are artificially high. There was an issue in detecting the onLoad event correctly and in some cases it would include a bunch of work done in onload handlers (amazon.com is a great example but it looks like your QVC test was impacted as well).
All fixed now but something to be aware of.
Mar 23, 2011 @ 16:52:08
Ah, good to know. I ran the tests yesterday, so yep, that would explain the QVC test. Thanks for the heads up, Pat.
How You Doin? 5 Free Ways to Check Your Site Performance « Get Elastic Ecommerce Blog
Mar 23, 2011 @ 19:49:50
[...] Test by location and browser version (IE7 or IE8). This is important as the farther away a user is from your servers, the slower the page load (this can be improved with site optimization tools). One request to WebPageTest, don’t forget us Macs! UPDATE – WebPageTest.org now supports testing for Chrome. [...]
Mar 24, 2011 @ 09:34:36
I’d love to see a row of averages across all the sites added to that table. Great post.
5 Free Ways to Check Your Site Performance | Web Design | Steamboat Springs, Colorado
Mar 25, 2011 @ 06:34:43
[...] tools). One request to WebPageTest, don’t forget us Macs! UPDATE – WebPageTest.org now supports testing for Chrome. top Comments are closed. Archives Select Month March 2011 (6) February [...]
Are you focusing on first and repeat views? Then you may be ignoring 96% of your page traffic. — Web Performance Today
May 10, 2011 @ 21:11:41
[...] Experimental new Webpagetest feature lets you test sites on Chrome [...]