Tim Ash, president of SiteTuners, a landing page optimization company, said one of the big challenges marketers face in optimizing Web sites and landing pages involves “a kind of fear of math.” He said marketers need to get past that and test different versions of landing pages, even if they do simple A/B testing.
“Landing page testing is based on statistics,” Ash said. “People are afraid to do it because of that or they will make novice mistakes like drawing conclusions from insufficient data.” Ash said that when it comes to landing page optimization, it may take weeks or months to gather a sufficient data set.
“If you had one version that was converting at double the rate of another, then it would be clear which was the winning landing page,” he said. “But if it's a 2% difference, it'll take longer to know that that's a real difference and not just an anomaly.”
His advice? “Start small. If you have a lot of traffic, say, "Hey, Mr. Webmaster, give me 10% of that traffic' for a simple A/B split test and change one feature, like the headline of your site. That way you don't have to bet the farm.” M