Why Contextual Advertising May Not Work For Video
from the square-peg-in-a-round-hole dept
There's been an explosion of interest in online video, not only from the likes of YouTube, but from established media companies, like Disney. Of course, there's always that prickly question of how to go about monetizing it. While some think that consumers will be willing to pay to download shows, some are banking on advertising, which would seem like the best option for the more viral type of online video. One startup, Dave.TV, is hoping to apply Adsense-like contextual advertising to video. For example, using speech recognition, they would insert an ad about computers if computers were being discussed on the video. While contextual advertising works very well in search, it still hasn't been shown to work nearly as well in text (when people are reading where they are, instead of searching to go somewhere else). Taking the same approach in video may be even more problematic. Imagine trying to serve ads for an episode of Lost based on what the characters are talking about. You'd get a bunch of ads for deserted islands. The current system, whereby ads are shown for cars, movies, and future sporting events would probably work a lot better. To understand the problem with contextual advertising in its current form on the web, consider a website for digital photography enthusiasts; text ads for digital cameras aren't likely to work very well since all the readers probably have them. On the other end of the spectrum, contextual ads aren't ideal for general news sites either -- an ad for Burger King isn't likely to play well when you're reading an article about a shooting at said restaurant. While ad-subsidized online content may make a lot of sense, simply trying to apply the same model across different kinds of content is a bad idea.