Top Low Hanging Fruit to Performance Optimize your Web Site and boost Business
Web Page Performance was one of the big topics at the Ajax Experience in Boston this week. Steve Souders - author of High Performance Web Sites and Even Faster Web Sites - talked about the top things to look into when analyzing web page performance.
Business impact on Web Site Performance
Steve presented some recently published studies from Google, Microsoft and Shopzilla showing how much of a business impact web page performance has. Google and Bing presented a study showing how many people leave the site with degrading page performance and many fewer clicks these users make. Lost users and fewer clicks lead to less generated revenue with their ads.
Shopzilla showed how they could boost their business and save on the hardware side by improving their web site performance:
Top Low Hanging Fruits
Too many network roundtrips due to too many embedded resources
Analyze the actual network roundtrips when requesting a page. Not all embedded resources (like images, css, …) get loaded in parallel by the browser. Each browser has a set of physical network connections that are used to downloaded embedded objects. The following dynaTrace network view shows that – when accessing http://280slides.com/Editor – 172 .png images have to be loaded by the browser. It also shows how loading these objects gets deferred due to the network connection limitation of the browser.
You can use several different techniques to lower the amount of network roundtrips. If images are static think about client side caching. Or consider CSS Sprites to reduce the number of images by merging them into one bigger image. Another technique is Domain Sharding to overcome the network connection limitations.
No optimal use of Caching and Compression
Analyze the embedded objects for the usage of client side caching and compression. Static images can be cached on the browser to reduce network roundtrips on subsequent page requests. Look at the http response to analyze current cache settings and analyze if they could be cached in case they are not. Same thing holds true for compression. Large files can be compressed lowering the required network bandwidth.
Check out this blog entry by Puneet Sangal that gives guidance on content caching and compression.
Failing network requests
Any other low hanging fruits?
Books like those from Steve Souders or blog entries from the web performance community cover many approaches and checklists to speed up your web page. Please share your experience and post your feedback on this blog.