What's Wrong with this Picture? - measuring effectiveness & value of brand integration and placements

Traditional ads (e.g. 30 sec ad on TV) have well defined metrics for measuring ad effectiveness and value such as ratings for a show (e.g. GRP), duration, and ad rate (based off CPM and ratings) etc.
These metrics cover the world of tradtional ads on TV and movies fairly well. E.g. If I know rating (which imply a % of households) for a show, I've a decent idea on how many households are seeing my ad at any time. Of course, with the coming of DVR (30% penetration by 2008), the effectiveness of these metrics is called into question as well

Now try to place the same metrics in the world of product placements & brand integrations and we run into issues:

1. Ratings denote the % population who are watching a show at any time. We expect them to be fully immersed in its contents. So, when an ad comes up, they are watching it. But this doesn't hold true for brand integrations. The integrations are inserted into show and are part of its subconscious character. E.g. Tom Cruise uses Land Rover's discovery SUV in MI-III in the Shanghai shoot out scene; I know the number of people who watched MI-III in the first week but what is the impact of Discovery SUV on them? I cannot get this data point by just looking at the ratings.
2. Ad rate. This is directly calculated based on ratings (show's popularity). E.g. American Idol, the highest rated show for 2006 till date, garners $1.3M per 30 sec ad spot for its finale. How do you do the same for brand integrations? Isn't a dialog mention with lead actor more valuable than a background?
3. Look at the web world. Google has made it a science of marketing, monetizing, and measuring ad placements. Everyone can take part and all are happy as they are paying for the ad on a pay per click basis, i.e. performance driven.

Here is a real example (I'll not name the brand to protect the innocent) - a brand had the highest exposure (number of seconds in a show) for a top rated show in 2004 season. Going by this data, this brand should have been the most effective for that show. But no one had ever heard of it, in fact brands far lower than it in terms of exposure were associated more closely with the show. So, what went wrong here? The brand was in the wrong places in the show, e.g in the background v/s with lead actor, etc. I don't know how much this brand paid to get in the show but they sure didn't get what they had hoped for. I wonder if they could do it differently, similar to the web world....

So, how we go about addressing these issues?

We need to rethink the way we measure effectiveness in the brand integration and product placement world. These parameters can be based off, but not the same as, the 30 sec ad spot (for TV). New techniques, based on cognitive science, need to be applied here.

Additionally, we need to rethink if techniques, which are successful in the web world can be applied here.

I'll lay out a proposal for a new technique for product placements and brand integrations in the next post.