Maintaining a decent website speed while also plugging in all the necessary analytics, adverts and product promotions is no mean feat.
And losing even a few seconds on your load time can have an adverse impact on usability, conversions and SEO.
In fact, stats included in our recent post on how to improve site speed show that the average consumer expects sites to load in two seconds, and slow loading websites could cost retailers as much as £1.73bn in lost sales each year.
But help is at hand. Google has a free PageSpeed Insights tool that anyone can use to measure the load time of a particular website.
A score is given out of 100 with the lower the score meaning the more room there is for improvement. It also analyses the content and gives tips and advice on what can be done to make that page faster.
Using this tool I thought it would be interesting to see what Google thinks of the homepage and a product page for each of the top 10 UK retailers.
Here are the results…
Amazon’s homepage score is 90 out of 100 and Google didn’t flag up a single high priority suggestion.
There were several medium priority suggestions though, including minimizing redirects, combining images into CSS sprites, serving resources from a consistent URL and optimising images.
Despite being littered with more product information, promotions, cross sells and consumer reviews than you could ever hope to read, Amazon’s Kindle e-reader product page achieved a score of 91.
However it did come up with two high priority suggestions: serve resources from a consistent URL and combine images into CSS sprites.
Argos’ homepage scored a relatively low 78 out of 100, which it might need to improve if it wants to become a ‘digital first’ retailer.
Argos’ product page for an Amazon Kindle e-reader scored just 59 with a number of high priority suggestions.
To speed things up it needs to leverage browser caching, optimise images, combine images into CSS sprites and serve resources from a consistent URL.
Apple’s homepage achieved a score of just 62, which is one of the lowest on this list. This is probably partly attributable to the large, hi-res iPad images that appear.
The high priority suggestions were to serve scaled images and leverage browser caching.
The iPad 2 product page scored a more respectable 76 out of 100, with the high priority suggestions being to leverage browser caching and combine images into CSS sprites.
Tesco’s homepage scored a rather rapid 95 out of 100, which is probably partly due to the fact that it has a limited number of product images.
There were no medium or high priority suggestions, so bravo to Tesco.
The Kindle e-reader page scored 79 and had a number of high priority suggestions: optimize images, leverage browser caching, serve resources from a consistent URL and combine images into CSS sprites.
Next scored a rather slow 23 and has a lot of points that it needs to work on according to Google. But while this score is much lower than the other brands, it didn’t seem noticably slower to load than Amazon.
Next’s product pages are even slower, scoring just 22. This is despite the fact that the page design is fairly sparse.
Play scored a respectable 84 out of 100 despite the fact that it has a large homepage with a number of images.
The product page for The Lion King DVD scored 85, which is good compared to many of the other pages on this list.
It has three high priority suggestions, including minimizing redirects, combining images into CSS sprites and leveraging browser caching.
M&S’s homepage scored 81 with high priority recommendations for leveraging browser caching, combining images into CSS sprites and minimizing redirects.
M&S achieves a respectable 81 out of 100 for its product page, but Google highlights several high priority issues.
But on the plus side there are no medium priority recommendations.
John Lewis scored a not-too-shabby 81, but again there are a number of high priority recommendations.
The Armani Jeans product page achieved a score of 76, and there were a number of familiar recommendations from Google.
As a high priority John Lewis needs to leverage browser caching, combine images into CSS sprites, specify a cache validator, enable compression and serve resources from a consistent URL.
ASOS automatically redirected me to the asos.com/men domain as I was logged into the site, but even then it still had a site speed score of 79.
ASOS’s product page is actually a lot quicker than the homepage according to Google’s tool.
The homepage scored 81 and there were no medium priority recommendations, however as a high priority it needs to leverage browser caching, combine images into CSS sprites and minimize redirects.
If you happen to be browsing for a pair of navy worsted trousers, then the product page you land on has a score of 73.
Google also had a number of high priority recommendations; leverage browser caching, combine images into CSS sprites, minimize redirects, serve resources from a consistent URL and enable compression.
Overall the standard was consistently high, with Next being the only retailer that achieved a score of less than 50.
It scored just 23 for its homepage and 22 for the product page I looked at, which was easily the worst score out of the 10 retailers.
Apple actually achieved the second lowest homepage score (62) and had the joint sixth lowest score (76) for its product page.
At the other end of the scale, Tesco clocked the highest homepage score of 95 and was the only retailer that didn’t get any medium or high priority recommendations on how to improve its site speed.
However Tesco’s product page scored 79, while Amazon had a more consistent score of 90 for its homepage and 91 for the Kindle product page.
In truth though, the differences in page speed between Amazon and Next weren’t that noticeable on the Econsultancy Wi-Fi connection, but it could be that the impact is more obvious for users with slower broadband.
However this test does give a useful benchmark for other businesses concerned about site speed, and shows that from Google’s point of view even major retailers can stand to make improvements to their page loading times.