Saturday, December 02, 2006

How I'd put advertising on user generated video

I've been thinking about video advertising on user generated videos for the past few weeks and this is how I'd do it.

One of the biggest challenges advertisers face on user generated video sites is the massive amount of videos whose quality and content varies greatly. If I was a brand conscious advertiser I would be worried about showing pre, intra, or post-roll to something that could be detrimental for my brand image.

My proposed solution is to use the number of views as a critical metric:

  1. Quality and content is reviewed at various thresholds of popularity. First level of popularity would be outsourced overseas, next level would be people looking for innuendos and subtitles that might be missed by non-native English speaker. Final level would be professional editors. At each level of review you'd also have more tags, etc added to the video to help further increase the velocity.
  2. Upon each review the video is tiered and classified which then drives the type of advertiser who can run on that video clip.
  3. As the views increase, start to place shift the insertion of the ad. When is first runs and there is no human review, run CPA and non-brand sensitive advertisers post-roll. As human review occurs then start then move to mid-stream. When it reaches Mentos/Diet Coke level of fame, pre-roll.

There are a few things that need to happen for this to work.

1. Need to revenue share with the poster or maybe their favorite charity. If there isn't a reason to keep the video on video sharing site, you'll have multiple versions running everywhere as the author puts the video on multiple sites to increase viewership.

2. Control the duplicates on your video site, use video IQ or something to reduce the copies. Not doing so will only reduce the velocity of the streams on a single video.

3. Keep copies of videos off other sites.

LMK your thoughts.

1 comments:

Rob said...

Based on some feedback, I edited this post to make it clearer. Thanks for the feedback.