Cracking the Online Video Monetization Nut: Adap.tv CEO Amir Ashkenazi (Part 8)

Friday, June 29, 2007 | 1 comment

Here I challenge Amir some more on the business strategy. Technology is wonderful, but are they solving the right business problem? Just because I am looking at a car chase in a movie segment doesn’t mean I am ready to jump and buy a BMW. There is little relevance. On the other hand, if I am searching for a BMW on Google, the keyword search is a direct evidence of intent.

I really wanted to get into his strategic level thinking to see how deeply he has thought through the issues he is going to face, and see if he has devised an effective strategy which he can build the entire company on.

SM: The reason I am not convinced about where we are going with this is because typically the best way to determine video relevancy is not necessarily recognition capabilities because video is often about entertainment. The real phenomenon that is of most importance is the segment that this video is watched by, and it has nothing to do with keywords, visuals and speech. AA: This is why the learning system is playing such a significant role in determining the ads displayed. Regardless of the genre, users respond to the ads, and we can use their response and the data generated by their response to select the right ad for specific content.

You are totally right – it is not all about content and context, there are many videos which are not informational. This is why all of the technologies I mentioned above are being used for a subset of the videos. At the end of the day those are all input data streams to the algorithms, and sometimes they are used and sometimes they are not. They are all determining the ad selection prior to having the movie live. Once the movie is live and we get user interaction, we learn from that and optimize the ad selection from that.

SM: I am going to re-ask this question in another way. I believe very strongly that the internet is evolving into Vertical Segments. Travel is a big vertical, real estate is a big vertical, and there are plenty of those verticals. When there are travel videos, the best place for a user to look for travel videos will be through some travel related site. What is your value proposition if we have already constrained the domain to travel? Are you then going to go for the destination viewers are interested in, because it seems like that would already be in the meta data? What are you bringing to the table that is new and extra? AA: Many times the meta data is noisy or non-existent when it comes to video. Second, given the fact you know the destination does not mean you know the right ad for that segment.

Hotels may work better than flights, and cars may work better than both. The only way to come up with the right ad is by experimentation. After running through several strategies you can see which one is more successful.

This segment is part 8 in a 10 part series Segment 9 →
Jump to part: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10

Bookmark It