It’s now official, Google are now including page speed in their search algorithm. For a long while studies have proved that Internet users much prefer fast loading websites, it seems now speeding up your website is not only going to make your visitors happy but may also have a small impact on your search engine rankings.
With that in mind I thought it provided us with a good opportunity to take a quick look at a number of popular e-commerce platforms to see how they perform on the page speed barometer.
Objective of the Speed Challenge
The objective of the speed challenge is to determine which hosted shopping cart platform had the best page load times when using a number of well-known site performance testing applications.
E-commerce Platforms Tested
For the purpose of this challenge we decided to only test a number of the more popular hosted shopping carts — BigCommerce, CoreCommerce, Shopify and Volusion. The reason for focusing on hosted e-commerce solutions was that each of these software providers has complete control over the web server environment and the underlying code base that the application is delivered on, which means if you are a customer you will be using exactly the same infrastructure as we are testing (hopefully making this test useful to you).
From each of the above providers we performed testing on a free trial account and a random live site that was obtained from testimonials/examples from their company website.
Homepage vs Product Page
We performed two separate tests, one on the homepage and the other a product page. The product pages of each application provider are very similar and may be a more realistic comparison than the homepage.
The home page layouts of each application tested were vastly different and caused fairly significant differences in results (particularly in the free trial account tests). For example the BigCommerce default homepage template consisted of thumbnail images of 12 products, while the Shopify and Volusion examples only had one.
Test 1 — WebWait
WebWait is a website where you can measure the speed of other websites. It ties precisely how long the website takes to load and render in your browser, including any images, style sheets, and JavaScript.
Test 2 — PingDom
PingDom measures for the load time of a website by mimicking the way a page loads in a web browser. The full-page test loads a complete HTML page including all objects (images, style sheets, Java scripts, RSS, Flash and frames/iframes).
Test 3 — WebPage.org PageTest
WebPage.org PageTest is an open source tool for measuring and analysing web page performance directly from your web browser. Developed by AOL it has been designed to automate load time measurement of a website and its objects.
Click to Enlarge
Test 4 — YSlow Firefox add-on (Yahoo!)
YSlow analyses webpages and suggest ways to improve their performance based on a set of rules for high-performance webpages. YSlow is a Firefox add-on that is integrated with Firebug.
Test 5 — Page Speed Firefox add-on (Google)
Google Page Speed is an open source Firefox/Firebug add-on that allows Webmasters and developers to evaluate the performance of their website. Page Speed runs several tests on a site’s server configurations and front-end code that are all based on a set of best practices that are known to enhance web page performance.
Combined Overall Ranking Results
When we finished the above tests we combined the score of each of the four pages tested to get an overall tally and then rated each shopping cart 1 — 4 based on the results. For example if you look at the Webpage.org test the for Shopify, the pages that were tested had the following results 2.762, 4.627, 2.353 and 4.627. Combining these load times gives a value of 14.369 which was clearly the lowest combined score.
Once each shopping cart had received a score for each test we combined the ratings total to come up with an overall ranking score, the lower the score the better. For example shopify received scores of 2, 1, 1, 1, 1 — which resulted in an overall total of 6.
I realise this is not the most scientific method of obtaining a result, but it at least use our challenge a winner while at the same time gives a valuable insight as to how you can conduct your own testing using these online tools.
Summary Of the Results
From the results of my testing Shopify comes through as the clear winner. BigCommerce, CoreCommerce and Volusion all finished pretty close together, but were well behind Shopify.
While none of the sites performed too badly, in my opinion the underlying reason that Shopify was able to stand out was that they were delivering a lot of the website elements via a Content Delivery Network (CDN). A CDN would have a huge impact on my testing as it was conducted in Australia. Somebody performing the same tests in the US would probably have a much different result as all of the e-commerce providers tested have web servers located in North America, while only Shopify is able to distribute content locally in Australia via their CDN.
Should you place a lot of weight on these results?
In my opinion, you shouldn’t use this test to base a decision on which e-commerce platform you should use. Your ultimate decision should come down to whether the e-commerce provider can give you access to the features and support that you need to be successful. While fast page load times are very important (in terms of usability and search engine rankings) make sure you do your own testing based on where your customers are located.
I would also anticipate that each of the e-commerce providers tested will be using Content Delivery Networks in the near future. This will dramatically help them close the gap on Shopify with their worldwide delivery times.
If anybody decides to conduct their own testing in a geographic location other than Australia please feel free to e-mail me your results and I will update this article.
{ 10 comments… read them below or add one }
Great post Dave! We have been reading a lot about this lately and it is really scaring me.
My site is very slow compared to your tests. I am weighing up the our IA recommendations v’s too much content. I am tweaking it hard at the moment and your post has really helped me shave some crutial seconds off already. Once again thanks again
Steve,
a lesson that I learnt a long time ago was to never make changes just for the sake of Google, your customers are much more important. In saying that, I really welcome this element being added to their algorithm as it ultimately promotes a better user experience through faster load times. It also rewards quality developers that are going the extra yard with best practices and clean code.
A question in regards to image file formats. Is one faster than the other? JPG’s v’s PNG’s??
That’s a good question and to be honest I wouldn’t have a clue, but I would have thought that at the end of the day it would come down to file size. Once again it is probably something that could make a very minute amount of difference and the time that you spend working it out would be much better spent on other initiatives.
Steve,a lesson that I learnt a long time ago was to never make changes just for the sake of Google, your customers are much more important. In saying that, I really welcome this element being added to their algorithm as it ultimately promotes a better user experience through faster load times. It also rewards quality developers that are going the extra yard with best practices and clean code.
+1
Dave,
Good info! Do you have any experience or information you could share regarding Americommerce.com? They seem to offer a lot of features; but as with BigCommerce, things aren’t always what they seem. Any insight would be appreciated.
Mary,
Unfortunately I am not familiar with Americommerce.com. I might take a quick look at the trial.
Having invested most of my time with BigCommerce/Interspire in recent times I haven’t paid too much attention to other options.
I can tell you speed tests are not all equal. For example if you have a site in shopify vs volusion. Lets say volusion has more images on the site you are testing or different features then you compare it to shopify which has less features and less images for the comparison. Then you say shopify is faster. That isn’t true though since it was not at all an apples to apples comparison. The only true measure of server speed is to put a blank html page up there and then test it.
Also running google page speed or yslow isnt accurate in this measurement for the same reasons. Let say someone on shopify is a good html coder and follows the appropriate guidelines. Then you have a someone on volusion who is a poor coder and has mega large images or many get request for images. You aren’t comparing apples to apples again.
Food for thought :_)
Thanks Nick you are 100% correct. At the end of the day everybody should do their own testing and this article was designed to show that there can be huge variances.
Adding a little more to my last message. For example on the Google page speed test
On CoreCommerce:
http://www.cierracandles.com/ 85/100
On Shopify
http://store.foofighters.com/ 73/100
On Volusion
http://www.thegreenloop.com 58/100
On BigCommerce
http://www.petrolfordogs.com/brands/Petrol.html 74/100
So as you can tell the test depends on which sites you use.