This article does not report reasoning part (i.e. why the growth rates are predicted as they are). So it is of limited value.

The problem of handpicking short-form YouTube videos to place advertise on is that of transaction cost. If they choose the automated (algorithm based) ad placing , the problem is the business practice and advertisers' mindset - the level of control over contents they would like to have must be low enough to choose and stick with that option.

Hosting video competition would work for some advertisers, but it could also be a very labor intensive process to review all the videos. They of course have an option to crowdsource the reviewing process, but then again, they have to be tolerant of unpleasant surprises.

I would say that people are not that awed so far by high-def TV, partly because they have not seen it, partly beacuse they are not interested in mere picture quality improvement. Higher speed data, however, means they can get some things done quicker, they can get less interruptions and errors. But here, too, I am not sure if the respondents fully know what is attributable to the speed and what is not. Still, it is an interesting knowledge. I wonder how their interests in mobile applications will compare with those. My guess is that even though the mobile apps. are new and potentially interesting to them, they might also be somewhat unfamiliar for people to want them strongly.