Are Video Banners a Good Idea?
For our website, we landed firmly on no, though it is a test we have planned - if it were to show a positive real world impact, we would certainly use it. I will update this page once we have a chance to run the test.There are a lot of downsides, and many (maybe even most) of our competitors are doing it - while it could be argued that we should match or exceed them, we felt that being different was more important than following a new trend.
More than likely though, your competitors don't use this sort of video banner, so you will be most different by going for it. Having said that, downsides still exist.
Bandwidth
A fairly short and compressed video can still be 10+ MB. On a good internet connection, this is fine. For more limited connections, we can degrade or fail in a few nice ways, but the result will always be worse than if it were not implemented at all. The first thing we can do is try to serve smaller versions to mobile devices. This may not help enough, however, and doesn't help with low bandwidth desktops and laptops. We can also just be sure to have a good 'poster', which will take the place of the video until it loads. Finally, we can just not show a video at all on small screens. With some combination of these, we can do alright for all visitors.The second half of bandwidth is the cost on our end. The cost is incrementally quite low, but can add up a lot for a site with high traffic. On our servers, bandwidth is a bit over 10 cents/GB, or around 1/10 to 1/5 of one cent per view of this sort of video. For a small site, that's just a few cents per day, but for one with thousands of views, it's easily a hundred plus dollars a month.