This is part three in a series of four posts on understanding the realities and challenges of split testing to improve your website’s conversion rate. If you missed earlier parts in the series, you can read them here:
Can you measure the results of your test? Does your test’s performance metric(s) matter?
In addition to limitations in creating and displaying test variations, you also need to make sure you can identify meaningful metrics and measure them accurately.
Generally speaking, the further your test is from the point of transaction, the more difficult it is to identify and isolate meaningful, reliable performance metrics. Offer pages, product pages and shopping carts are the most commonly tested because it’s easier to measure the results and the results usually relate directly to revenue or leads without subsequent influence from external factors.
Once you move out to the homepage, category pages (eCommerce) or information pages, your most reliable metrics become intermediate metrics, usually click-through rate to a desired page or pages. You should still measure sales or leads, but I’m not out to confuse anybody, so I’ll forgo the explanation of this until another time. Suffice it to say things get more complicated, so lean mostly on your most immediate metric for tests like this.
Despite these challenges, you shouldn’t be discouraged from venturing into split testing. In fact, now that you understand the challenges, your chances of success go way up. Go forth and realize the full potential of your website.