Robert Scoble writes about Internet Video Business Challenges explaining some of the economic problems with internet video. He uses some assumptions that video can generate $10 per 1,000 viewers in revenue, bandwidth costs $0.14 per GB, his average video file is 200MB, and that that it costs $210 in fixed costs just to make video.

Based on these numbers, the business model just doesn’t work. You start out spending $210 and then lose $18 per 1,000 ($10-$.14(1000*.2GB) people that download your video. If that were the case nobody would be doing internet video.

But that is flawed. If you do the math right, video works.

From what I know, radio CPM is about $5 ($5 per 1,000 listener impressions) and video is closer to $30-40 per 1,000 impressions. A 200MB file should have the ability to support 20 minutes of video and if it were TV, 10 minutes of ads or about 20 ads. That would be about 1 ad per 1 minute of video. Maybe internet TV can’t handle that ad load, but 2 minutes or 4 ads per 20 minutes of video seems reasonable. So if we get $30 per 1000 impressions a video should should be able to generate $120 per 1,000 viewers. At that rate, the show is generating $92 per 1,000 viewers in contribution to fixed costs. A break even of around 2,300 viewers per show.

I think Mr. Scoble should be able to do that. And if not, he can always cut his bandwidth costs by using a service like Redswoosh.net.

With a service like Redswoosh, he pays $0.03 per user per month, not total bandwidth per month. So if he produces 1 show per day and he has 1,000 users he would pay $30 for the entire month worth of video distribution or about $1 per day. (I guess the pundits were right when they said bandwidth would be free)

While some of my assumptions may be flawed, at least the numbers are directionally correct. I’m surprised Mr. Scoble with all his video blogging experience doesn’t see it that way.