YouTube has announced that they plan on using a Google-like ad model where sponsored videos will appear as part of search results. Some quick reactions:

• I’m surprised they didn’t do this before. This is pretty much a replication of the Google model and, given that Google does own them this is pretty much a layup.
• I’ll be interested to see who ends up buying really popular searches on YouTube. For example will Toyota be buying “cute kitty” just for reach purposes?

The big question for me is, what are the value-added services YouTube is thinking of here? They already have sponsored pages of sorts, and these ads should go a pretty significant distance in driving traffic to those pages. Also, they know quite a bit about their consumers; will they be able to target these ads on layers outside of search? For example, will I see a different ad for a movie trailer than my wife? Perhaps my trailer will have a lot of explosions and ninjas and hers will have a couple dreamily looking into each other’s eyes? This would sort of be a type of multi-dimensional targeting.  This type of targeting would have planning applications as well, much like Google AdPlanner.

Also, are we going to see “YouTube” studios - an online facility that will allow advertisers to make templated videos that will feel like a YouTube video, rather than the 30 second spot they just ran on TV? Finally, are we going to see a full integration of YouTube into Google video? This seems like the first logical step.

Lots of questions. Not many answers. I’m looking forward to seeing how this develops.

I just spent a lovely two weeks in the northern wilds with my family.  No Internet and, aghast, no cell phone.  We did spend a bit of time in the airport however, which led me to some thinking about the direction of the Internet and the importance of the current incarnation of search.  But first a story:

My wife and I were sitting the middle of an AirTran flight and I was looking up at first class (yes, they do have 1^st class in AirTran flights - news to me). I saw a distinguished looking older gentleman whom I recognized, but didn’t know where from.  As soon as we got off the flight in Burlington, VT the gentleman was talking to a news crew.  Still, I recognized him, but had no idea from where.  My first thought was that when I got home I would Google him.  But alas, there is no way to Google “older guy I recognized on flight to Burlington” and have it be remotely meaningful.  Since I appear to have outsourced my ability to remember things to the Internet, I never got my answer.

This is because, as we all know, Google, and most search for that matter, looks for words.  We’ve found work-arounds with tag clouds and other forms of meta-data, but no real answer to the root problem.

Why should we care now? Well total time spent watching Internet video us up more than 30% over 6 months according to VideoCensus, versus, basically flat overall Internet time.  The image based (non-text) Internet is growing.  Given that we are still lacking (good/real/useful) meta-data standards for video we have a problem.

We can’t search it.

I think this might be a problem.