A/B and Multivariate Testing [SIMPLE GUIDE]

The migration of traditional commerce onto the Internet is accelerating. In 2010, e-commerce sales increased by 14.8 percent, according to the United States Department of Commerce. E-commerce is a growing field, marked by technological innovations. The ability to sell online is revolutionizing sectors of the retail industry. Companies like BestBuy, Target and Kohl’s are now risking losing market share to online retail giant Amazon.com. The data from the Commerce Department also showed that e-commerce sales rose faster than traditional retail sales for 2010. Clearly, e-commerce is on an upward trend.

Websites and businesses that wish to ride the e-commerce wave must improve their approach. Much like a physical store, a website is designed to convert visitors into customers. E-commerce relies upon the strategies and techniques of Internet marketing to improve the rate at which visitors converted into customers. The conversion rate, as this is known, is the most critical statistic for a website to track. The conversion rate can spell the difference between success and failure. Even if a website’s traffic volume is not very significant, a high conversion rate can turn this low volume into large profits.

Websites improve their conversion rates by testing different aspects of their pages. Business owners can avail themselves of two different forms of testing: A/B and multivariate. A/B testing takes one page element at a time and uses two different versions to see whether visitors like one or the other. Multivariate testing does the same thing but with multiple page elements simultaneously. The different page elements include images, content, titles and purchase buttons. The placement, size and variation of the images can all be tested, as can the placement, font, and combination of the content with different images or titles.

A/B Testing

A/B testing can be implemented in two ways. The website owner can either change the element of the page as it is loading or redirect it to another page. The former method is good for testing a single element while the latter works best when changing an entire layout. It is critical to select a metric that will define success. This metric depends on what the goal is: Making a purchase, signing up for a newsletter or joining an email list. Measuring how many visitors achieve one of these goals will give the business owner a measure of the success of his website. It is important to not let emotions override the results of the test, which can be surprising or counterintuitive. A/B testing is usually the starting point for testing and is more suitable for smaller sites and blogs as it can be carried out with less traffic than multivariate testing.

Multivariate Testing

Unlike A/B testing, a multivariate test allows an owner to create multiple versions of the same elements. Each version is shown to a different random set of visitors and the results are measured. With multivariate testing, it is critical not to take on too many challenges at once. Focusing on a few sets of elements at a time can prevent a lot of headaches later on. Every test to be run must have a defined goal in order to be successful. Otherwise, a mass of useless data will be generated that tells the website owner nothing. With multivariate testing, it is critical not to take on too many challenges at once and the more powerful the multivariate data analysis software the better.

Advantages And Methods

Business owners can take advantage of websites that assist in the testing process like Google Website Optimizer (they have some great videos and tutorials on how to get started)which is free to use and Visual Website Optimizer. Online businesses can gain a huge advantage over their competitors by using A/B and multivariate testing. The e-commerce revolution will continue as these testing methods become more widespread.