How effective are event sponsorships? (Part 2)

I was just pondering the effectiveness of event sponsorships last week, and today I came across this article(via Google Alerts) about Allstate’s brand exposure from NASCAR. A sponsorship measurement firm analyzed almost 600 sponsors of NASCAR to find the value of their sponsorships.

“Primary and secondary car and driver partners were analyzed, along with all race venue signage, and the myriad graphics and audio mentions from the races’ TV broadcasts… Each of those individual detections was then evaluated based on its duration, average size, location and relative isolation (or lack thereof) from competing brands.”

Based on this evaluation, “Allstate’s 702 detections generated $32.9 million in exposure over the course of the season, putting Allstate No. 9 overall among the nearly 600 companies tracked.” The dollar amount was determined using an ad-value calculation.

This is one way to measure sponsorship effectiveness, but it doesn’t seem all encompassing to me. Shouldn’t measurement, in some way, include overall sales? Just because people saw the signage on TV doesn’t translate into higher sales.

In addition, I wonder if there’s a difference in impact on the audience based on the company. Are unpredictable sponsorships more effective? Auto manufacturers and insurance companies make sense with NASCAR, but does Viagra and Claritin get more attention from viewers because it’s some-what out of place?