I posted a few weeks ago (Optimizing your site) about how I was running some tests with Google’s new free Website Optimizer tool. This tool lets you perform multivariate tests on your website, so that you can test the effect of changing multiple items at once. Google manages all of the hard stuff, letting you focus on watching the results.
I ran a test on the primary site I manage, www.lcdarms.com. There is a billboard on the homepage. Some time ago, we started placing a flash element in the billboard space. Nothing fancy, not once of those annoying 5-minute movies. Just some movement and splash of color to attract your attention to the graphic which finally settles there. My thinking was that not only would this dress up the homepage, it would also improve clickthrough to whatever item I was featuring there.
According to the test I ran…. boy, was I wrong! I tested the flash element vs. a simple jpg. The jpg was identical to the final image which the flash element would resolve to, so there were no hijinks there. Over the course of 2 separate tests I’ve run, (and continuing with a 3rd test in progress now), each time testing a flash element vs. a jpg, the simple jpg image has far improved the click through. We’re talking about clickthrough increases of 50%!
To me, that’s really counter-intuitive, and not what I ever would have expected to see. In the interest of science, I’ve formed a hypothesis explaining the results:
- People who are web surfing don’t think about the technology behind a website. They don’t think ‘ooh, it’s flash’. They look at a site and see it as a cohesive whole.
- People have been conditioned to think that moving graphics denote advertisements. And as we know, they tend to ignore advertisements. ‘Banner blindness’, as it’s been called.
- Thus, when people come to a website and see something moving, instead of attracting their attention, it diverts their attention to other static items. Nevermind that in this case the moving item only has a motion for a second or two and then settles on a static image. They’re already blind to it and they don’t go back.
I have some further test ideas in mind which I’m going to use to explore this theory further. More on that in a few weeks.