A 19.53-second page load, Keynote finds, but much less for returning visitors.
Bill Siwicki , Editor, Mobile
The popular phrase “Go big or go home” does not apply to crafting mobile commerce web site pages. The bigger the page, the more time it takes to load. The more time it takes to load, the more frustrated consumers become, which can lead to consumers abandoning the site.
E-retailer GameStop has a massive mobile commerce web site home page: nearly 50 page objects, such as images and boxes of text; a weight of between 500-600 kilobytes; and the need to fetch page content from 10 unique web domains, says mobile and web performance management firm Keynote. An optimal size for an m-commerce site page would be no more than 20 objects, though 10 or fewer is ideal; between 100-150 kilobytes in weight, anything more is overkill; and accessing content from between 6-10 domains, anything above 15 is unacceptable, Keynote advises.
For the week ending Feb. 24, it took on average a whopping 19.53 seconds to load GameStop’s m-commerce site home page, and the page loaded completely and successfully 98.99% of the time, according to the Keynote Mobile Commerce Performance Index. Weighting and combining the load time and success rate earns GameStop an index score of 556 out of 1,000. The average score for the week among the 30 retailers on the index was 768 and the high score was 954.
“GameStop is an interesting case in that it appears they have decided to focus nearly all their optimization efforts on repeat visitors, rather than first-time visitors,” says Venkatesh Giri, mobile performance expert at Keynote.
The m-commerce site home page loads on average for repeat visitors in 2-3 seconds, Giri says. GameStop has implemented an HTML5 caching feature called cache manifest, Giri explains. Cache manifest is a simple text file that lists the resources a browser should cache for offline access.
“The advantages of using cache manifest are multiple, including: offline browsing, which allows a user to navigate and access the web site without wireless connectivity; increased speed, cached resources are on local devices which load faster; and reduced server load, the browser will only download changed resources from the server,” Giri says. “Using cache manifest, GameStop provides a great user experience and faster page load times for their return users. The return user will download less than 10 objects when they revisit GameStop.com because most of the nearly 50 objects are already cached on the device.”
GameStop can improve its mobile performance by using image optimization techniques, Giri says. The e-retailer can use CSS sprites, a web programming technique that enables multiple images to be saved as one, thus reducing the number of web server calls required. And it can use data URI, which stands for universal resource identifier, a string of characters used to identify a web resource such as an image. Data URI is a scheme of encoding data within a web page that make up page elements such as images or Cascading Style Sheets, or CSS, a mark-up language used to define pages and denote where elements appear on a page. With multiple elements encoded within a page, no extra HTTP server request is made to fetch the embedded elements as opposed to a request for each element.
“GameStop can climb up the Keynote Mobile Commerce Performance Index once they figure out image optimization and combine images to reduce HTTP server requests,” Giri says. “And overall they need to adopt a more balanced strategy that is focused on both first-time and repeat visitors.”
GameStop, No. 90 in the Internet Retailer Mobile 400, says the m-commerce site as it exists today is a short-term, interim step to a faster site.
"We agree that it can provide a sub-optimal experience for some first-time visitors and are actively working to address that," says Ashley Sheetz, GameStop vice president of marketing and strategy. "We are in the process of developing a mobile solution that will address the load issue and will provide a more seamless and engaging experience for our customers."
Toolfetch.com LLC topped the index with a load time of 3.12 seconds and a success rate of 99.57% for a score of 954. Barnes & Noble came in second with a load time of 5.33 seconds and a success rate of 99.89% for a score of 943. And W.W. Grainger Inc. came in third with a load time of 4.15 seconds and a success rate of 99.36% for a score of 910.
Toolfetch.com is No. 248 in the Internet Retailer Mobile 400, BarnesandNoble.com Inc. is No. 31 and Grainger is No. 76.
Click here and then click on Keynote Mobile Commerce Performance Index Part 1 and Part 2 to see this week’s complete results for all 30 retailers on the index.
Keynote Systems measures 30 representative m-commerce sites exclusively for Internet Retailer. The sites include merchants in multiple categories and channels, and of multiple sizes, ranging from such giants as Amazon.com Inc. to midsized retailers like Toolfetch.com LLC. Keynote tests the sites in the index every hour Monday through Sunday from 8 a.m. through midnight Eastern time, emulating three different smartphones on three different wireless networks: Apple Inc.’s iPhone 4 on AT&T, the HTC Evo on Sprint and the Droid X on Verizon. The HTC Evo and the Droid X run Google Inc.’s Android operating system. Keynote runs the tests in New York and San Francisco.
Keynote combines a site’s load time and success rate, equally weighted, into a single score. Given that both performance and availability are important, the score reflects the overall quality of the home page; a higher score indicates better performance. Scores also reflect how close sites are to each other in overall quality. The index average score is the midpoint among all the sites’ scores.
Keynote is ranked No. 2 among web performance monitoring firms in Internet Retailer’s Leading Vendors to the Top 1000 E-retailers guide.