up

Zoompf's Web Performance Blog

Note: Archived Content

This is the archived version of the Zoompf blog. Since our acquisition by Rigor, all our new research and posts on web performance are being published on The Rigor Blog

9 Top Online Retailers Race in the First Web Performance Derby

 Mark Isham on May 23, 2013. Category: Alexa Research

If you’ve never checked out WebPageTest, you should. It’s a fantastic open source web performance analysis tool supported primarily by Google.

Here They Come (c) Dave Black www.daveblackphotography.com

One of the great new features of WebPageTest is the ability to capture video showing how long it takes your site to load relative to your competition. This is a great way to objectively see your site’s user experience – and as we all know, the slower your site, the more likely your users will look elsewhere.

Towards this end, I thought it might be interesting to generate a video comparing the home page load times of 9 of the top online US retailers.  A race of champions of sorts. In fact, watching the video feels much like watching a horse race: Win, place or show – how will your pick finish?

So to take this weak metaphor and run it into the ground (or is that “beat a dead horse?”), here are my top 9 picks, favored by Alexa rank:

  • Amazon (7)
  • ebay (21)
  • Walmart (186)
  • Etsy (191)
  • Target (335)
  • IKEA (342)
  • Best Buy (394)
  • Home Depot (403)
  • Dell (434)

To set this all up, I visited webpagetest.org and configured these sites on on the “Visual Comparison” tab as shown below:

WebPageTest

Super easy. This in turn configures a near simultaneous race of all 9 sites from an instance in Dulles, VA running on IE9. After running the test, you can select various viewing options (more on that later), and export the video to a file. For the first race, i chose to complete the race on “Visual Complete” (all visual elements loaded), as you can see in this video:

Went by a little too quick? Don’t worry, you can also export in slow motion for a more dramatic effect as shown here:

And the winner is…

IKEA

…with a smoking fast 1.1 seconds! Nice work guys!

Much like their furniture, IKEA’s site owes much to simple and functional design. Fortunately their website does not require you to follow a 20 page instruction manual to put together…

So the top rankings are:

  1. WIN: IKEA at 1.4 seconds
  2. PLACE: ETSY at 2.3 seconds
  3. SHOW: Dell at 3.5 seconds

So what happened to the odds favorite Amazon, coming in dead last at 11.3 seconds? Surely that can’t be right?

Well, the devil is in the details. While the fact that their site is “content heavy” no doubt contributed to some delay, this still doesn’t mesh with our expectations for the #1 retailer. So what gives?

Well, as mentioned above the test was set to finish with “visually complete” (all visual elements loaded). Amazon loads quite a bit of their content after “document complete” (when the user can start interacting with the page), so the results can be somewhat skewed. So what happens if we instead measure by “document complete”? Well let’s take a look here:

Amazon suddently swoops into a very respectable third place at 2.2 seconds! Quite a difference to be sure.

So why not just always measure “document complete”?

This speaks to one of the big questions in the web performance world – what is the best way to measure “speed”? Is it when the user can start interacting with the page, or when all the visual elements of the page are completely loaded? You could easily argue on both sides of this coin, and each website developer must make a conscious decision of how they want to play this card. Show less, but faster (IKEA) vs. show more, but interact sooner (Amazon). This is a design decision each with its pros and cons, but what’s great is WebPageTest allows you to measure either way to suit your prevailing philosophy.

So I hope this helped show off a really cool feature of WebPageTest. And of course, if you want to learn more about how to speed up your website, try a free performance scan using our own awesome tool at https://zoompf.com/free.

 

Comments

Have some thoughts, a comment, or some feedback? Talk to us on Twitter @zoompf or use our contact us form.