UK online performance scores are in
Wednesday January 22 2014
End-of-year retail site speed tests reveal winners and losers, while the cost of excessive third party tag usage on performance is called into question
The latest online shopping benchmark data from performance technology company Compuware has found that UK retail’s web responsiveness managed to keep pace with consumer expectations during the last holiday season.
Performance winners and losers
BizRate continued to top the November ranking with 0.582 seconds, but was overtaken by the superior consistency rates Tesco UK during the month ending 1 January. But Tesco only raced to its number one position in December from number 14 the previous month, with an average response time of 1.023s, consistency rate of 0.878s and availability of 99.86%. Its availability also rose to 100% in December.
The best performing fashion site during December was Next with an average 0.654s response time, Marks & Spencer rose from 38 in November to 14 with 100% availability but a 0.951s response time, and TopShop brought up the rear in 47th place with a 4.050s response time, 4.628s consistency and 99.59% availability.
Meanwhile a new study has also found many of the UK’s top retail and fashion websites saw their transactional response times more than doubled by an overabundance of advertising and marketing tags on each page during the festive trading period.
Third-party tags hampers mobile speeds
The third party tag findings, compiled by website performance specialists Intechnica
, were based on external monitoring data taken from iPads and Samsung Galaxy S4 devices over a 3G connection.
It compared the total response time of a land-and-search journey on 12 leading retail sites, against the same journey ignoring all third party tags (advertising, analytics, marketing etc.) during the key December and early January sales period.
Slow user experience is strongly linked to reduced conversion rates, page views and brand image. The industry benchmark for loading a single page is around 3s. As this study looked at a two-page journey, one of which being a more complex search result page, the benchmark was just below 10s. In reality, the average transaction took up to 20s to complete, yet without any third party tags the average would have been just over 10s.
“The real killer in terms of performance was the sheer number of external connections,” stated Larry Haig, performance monitoring expert at Intechnica. “Each connection adds latency, adding to the response time of the page. On average, these tags typically account for almost two thirds of connections on the pages.”
Putting bottom line sales at risk
“The disparity across different websites was astonishing: some sites we tested had less than 10 external connections, while some had over 200 and removing these from the site would have a dramatic effect on response times, page weight, and ultimately conversion rate on those pages.”
“Of course advertising and marketing tags are important to these businesses, but you have to ask which of these tags are necessary during peak trading times when bottom line sales should be a top priority.”
Sites with the most third party tags and external connections, such as New Look, were made significantly slower by the inclusion of third party tags, in this case accounting for 137 connections and 1.6Mb of the pages. The average transaction response time for the full site was a staggering 30s, yet by cutting out tags such as advertising and marketing, this would have been cut down to just 6s.
By comparison, Tesco’s two external tags only added three connections and 15Kb to their pages, resulting in an average transaction response time of 5.2s, adding just 0.2s to the transaction.
Tagged as: M-commerce | mobile | e-commerce | web | optimisation | speed | response | availability | performance | tags | tagging | Christmas | Compuware | Intechnica