How was the testing done? Every other pageview delivered over a 7 day period had the video?<p>If you tested by running one for 7 days and the other for 7 days, there are numerous other possible factors. Even the comparison of Friday versus Saturday traffic could have enough variation on a small sample to skew results. Internet congestion might have slowed the site down during one test, awareness might predispose people to be more likely to convert beyond trial or even the fact that the video version renders more quickly to the surfer proving that html design and page render time may have been the deciding factor. I didn't see any mention on the site how the trial was actually conducted, but, based on the other test results, I don't see enough isolation of variables to decidedly say the video was the sole factor.