YouTube Advertising

In a move that we knew was coming, YouTube have made finally unveiled their video ad platform. As expected its clearly a powerful and well executed play, but (so far at least) significantly less aggressive than it could have been.

Liz Gannes over at NewTeeVee has written a characteristically excellent summary of the program which can be viewed here but the main points are:

- Flat $20 CPM for ad impressions
- Targetting based on Location, Demographics [I'’m guessing where available from registered users], Time and Genre [Note: no tag-based targeting or keywords from voice recognition etc].
- Click-to-Play ads, based on slide-up overlay ad-previews occupying the bottom 20% of the video
- Stats on impressions, click-throughs, percentage of video watched, comments, and subscribers.

Although most of this is fairly unsurprising, and similar in many ways to what we'’re doing (with a lot less money spent on R&D I'’d wager!), I am intrigued that they'’ve chosen to base their pricing model on CPM for ad-preview impressions, rather than per-viewer who actually watches the ad itself. Based on the click through rates I expect they'’ll see (probably about 4-5%), a $20 ad spend will yield about 40-50 views, working out at $0.45 per view. But if your ad gets lower CTRs, then your PPV will rise rapidly. Certainly once advertisers work this out (and it won'’t take long), LiveRail’’s model will be a lot more attractive. With LiveRail advertisers can say they want to pay $0.45 for each view, and then know they'’ll only pay based on results; not just the opportunity to get results.

Google must believe that advertisers will prefer the predictability of a flat CPM basis for total spend but uncertain cost per view, rather than a less certain total spend, but very predictable cost per view. Since video ads tend not to be bought by the end customer directly, but through their agency or media buyer, the simplicity of saying “we bought 50,000 ad-previews and it cost $1000″ might be more attractive than saying “We set a daily budget of $1000 and it yielded approximately 2,500 views of the full ad”. From a publisher perspective, the stability of knowing you'’ll earn X% of $20 per 1000 views may also appeal, but the lack of flexibility for high desirable content/viewers means that most premium publishers will lose out. Google have not revealed what that magic revenue share rate is, but I'’ll be keeping my ear to the ground on that one - I suspect if will be variable to attract the more valuable publishers to the site.

Quite how this all pans out remains to be seen. If it turns out advertisers prefer to pay for their ad preview to be shown, and worry about CTR later, then we can adapt to this - it would be very easy to us to integrate, but I wouldn'’t be at all surprised if we see Google moving towards LiveRail’’s pricing model rather than the other way around.

The next concern is whether Google plan to split this advertising system from YouTube, and open it to be an open-ad platform serving video ads into third-party publishing platforms. As long as the two are connected, big name advertisers will remain wary (due to the risks of being associated with unsuitable UGC and/or copyright-infringing material) and their volume will be limited by the number of publishers willing to hand over their content to YouTubes’’s control.

So far it looks like this like Google have some interesting ideas, and this could be the ideal solution to monetize the YouTube site, but they are far from positioning themselves on top of LiveRail’’s market as a wider-scale advertising platform for video content online.