Ellen Julian

Website Testing, Page Load, and Foie Gras

Chances are that improving your website’s page-load speed is one of your marketing team’s top New Year’s resolutions. It’s well-documented that slow-loading pages affect customer satisfaction, SEO, and conversions, but marketers continue to make the same mistake every year by unintentionally slowing down their sites with heavy images, video, and website code. What’s an optimization team to do to break this cycle and teach marketers that they are creating the equivalent of website foie gras–a rich and appealing user experience that is detrimental to your brand’s health? Here are some suggestions:

Preview and measure the impact of each variation. SiteSpect product manager Iwo Kadziela suggests adding page load metrics to all of your A/B test campaigns. Customers typically measure how variations affect page speed on all of their domains -- mobile and ecommerce alike. Concrete metrics that document how much fat is contained in nice-looking but weighty web pages ensure that your marketing team gets the message on how they are slowing down the buying experience.

Optimize mobile performance with responsive design. Kadziela has also found that being able to target a mobile device server side allows you to detect and optimize requests. Targeting lets you identify mobile users and customize the path to an image before the browser starts reading and rendering the page, ensuring that mobile users get a smaller version of an image in a single HTTP request.

Combine this technique with an automated performance optimization solution to further compress and optimize images and other resources on your site. Such optimization not only improves speed and the user experience, but also lifts conversion and engagement.

Put your website on a page tag diet. Finally, limit your use of fattening page tags. SiteSpect examined how five JavaScript-based testing and targeting tools introduce latency and other visual anomalies into the user experience. We found that these JavaScript-based tools affect performance in all cases, ranging from 7% - 31% degradation in speed. Intuit Payroll’s web marketing team learned this first-hand as part of a web performance initiative. Following site improvements based on a performance audit, Intuit found that it did not need legacy page tags and when those tags were removed, page load times improved by an average of 35%.

So here’s to starting off the New Year right by helping your marketing team achieve its 2015 resolutions. And remember to take it easy on the foie gras!

Tags: A/B & Multivariate Testing Site Speed