Etailers must constantly improve their website to stay ahead of the pack, which is where AB testing comes in.
AB testing is a simple process that can give real insight into users’ behaviour on site.
It involves having two versions of an element on the site and testing which works best. Retailers can test a number of elements in this way, from the design or layout of the site to the checkout process. By measuring which version is the most successful, retailers can then apply this to the live site and improve the visitor journey.
AB testing is easy to do, often without the need of third-party involvement. So should all retailers be AB testing? Martin Newman, founder and chief executive of ecommerce consultancy Practicology, believes so: “Testing should be an ongoing occurrence. Most retailers don’t test too frequently, partly because they don’t always have the right people in the team doing the testing. But there’s no reason why other members of the ecommerce team can’t conduct AB testing using the likes of Google Website Optimiser, which is free and easy to implement. Testing is an obvious way to drive uplift in conversion and sales.”
Kristine Kirby, ecommerce and multichannel consultant, agrees but points out there are a number of common mistakes. “Retailers need to ensure they always have a test plan before they start. By being completely clear about goals, retailers can ensure their AB testing is effective. Campaigns should run for a number of weeks to gain sufficient data and only one campaign at one time so it is clear which change has caused the lift in sales, traffic or conversion.
With the quick pace and constant change of online retailing, AB testing should not be a one-off experiment but a constant process. Matt Connolly, founder of ecommerce consultancy Pollen House, says: “Things change and what’s right one week may not be the next. As for what you should be testing - start with the areas that have most riding on them financially, and ensure you have enough of a sample size.”