Attracting customers by advertising and price reductions is almost an unconscious tactic among traditional retailers. In the world of online retailing, website owners are intensely focused on getting more people to come to their sites. Where retailers with brick and mortar stores use old-fashioned techniques, online retailers use various online marketing techniques to attract visitors to their sites. This tactic only takes a website so far. Increased traffic flow will not necessarily mean increased sales.
The website owner must go a step further by actively trying convince these visitors to become become customers. Unlike a physical store, on the Internet this in person, but it can be achieved by changing a website so that it presents better and looks more attractive or professional. Trying to find out which element on a page is the key is impossible without running some sort of test. A/B testing, also known as split testing, is a simple technique that can methodically proceed through the elements on a site and discover what visitors like and dislike about each one.
How A/B Testing Works
Split testing works like this: A single item on a page, such as an image, button or layout, is changed. This process creates two different versions of the same page, an unaltered and an altered one. The unchanged version is used as a control to measure any difference the changed version has on the visitors. The aim of a split test is to discover whether the changed version has a higher conversion rate than the unchanged version. The conversion rate measures how many visitors are converting into customers. This is a crucial measure for the website owner because a high conversion rate sustains profit levels through varying traffic amounts.
Setting up the test is just as easy as running it. The only potentially problematic step is choosing and defining the goal, which will determine what components of the web pages will be tested. The goal is typically identical to the action the website owner wishes visitors to perform. For a website with a goal of convincing visitors to make purchases, the pages to be tested will include the checkout page and the product pages. Elements on these pages to include will be images, colors, footers, headers and call-to-action buttons.
Creating the test takes little effort, particularly if the website owner is using a free service like Google Website Optimizer. The only consideration to be taken into account is the kind of element to be tested. A page-wide factor like a theme or layout is better handled by redirecting visitors to an alternate page containing the changes. Conversely, a solitary item like a header or footer can be programmed to change as the page is loading. While the results are simple to understand, the changed page either has a higher conversion rate or it does not. If you want to read up further on methodology Maxymiser provider a guide to A/B testing and other types of testing.
Caveats
The simplicity of split testing conceals various ways the website owner can botch this technique. A common mistake is to surprise regular visitors by showing them changed versions of pages. This problem arises when testing product pages and altering elements such as prices, pictures or other essential details. A quality A/B testing program will have a method to discriminate between new and old visitors.
Another error is for the website owner to let his personal feelings override the test results. If the test is done properly, he should expect counterintuitive data to be produced. The whole point of testing a website is to discover what visitors actually like, not to confirm what the website owner thinks he already knows. This issue may arise when dealing with colors that seem to clash with one another, such as bright green contrasting with dark red or some other combination.
