I've long advocated a move to buying and selling something more determinate and tangible than a bag full of magic beans... uh, I mean, impressions.
The reason for this is that though the impression continues to serve as the 'atom' of the material advertising universe, it is not, ultimately, what marketers and their media should be concerned with.
It should be audience that marketers and media mavens alike should be concerned with; it should be individuals. And, in one way or another, it typically is. The problem is that we have traditionally been forced to use any number of proxies for audience rather than the audience itself. Like jets of radiation and light at the event horizon of a black hole, impressions 'stand in' for an individual's encounter with advertising, but it is not the individual him or herself.
Impressions, after all, do not exchange money for products and services-people do.
Tying ourselves up in debates about impressions is not ultimately going to lead us to the fluid but identifiable 'truth' of online media's relative value to other media. It is the individual and the role he or she plays as a constituent of an audience that should serve as all of media's normative means of valuation and NOT the availability of impressions.
Like I said, it takes persons to buy product. Impressions may be a valuable way to represent the 'volume' of communication delivery, but the impression's existence is incidental to that of the existence of a person. Therefore, audience must be our currency.
Advances being made in the industry to provide ratings are not a bad first step, taking us in the right direction towards ultimately settling on tangible (or even potential) audience. Though ratings, too, can be determined by impressions as components, the rating point keeps the media currency representative of audiences (the fact that impressions can't be made without people is accounted for). When dealing with online media, the relationship is really one of impressions to themselves.
Yes, impressions can't be generated without persons, but the counting of impressions currently doesn't account for that, and the buying of them does not come with assurances of discrete audience. Regarding online, when a tree falls in the forest and no one is around to hear it, the answer is that the tree doesn't really make a noise because the tree can't fall at all.
In television, determinations about reach can be discerned and do, indeed, inform the buying of the media. That is what ratings points allow. Reach (an estimate of audience) times Frequency (an average impression-per-person) equals Ratings. Online, I buy a million impressions and I have no guarantee from the media property whether I'm getting one incident of communication with a million people or 10,000 incidents of communication with 100 people.
I've used some of the R/F tools available out there: Telmar's and Atlas' GRP Forecaster. If you are to take the numbers on faith, the GRPs are small and, based on the CPMs currently in vogue, expensive. CPPs are often in line with what they are for network television. Do they deserve to be? If they were relatable to monetary reconciliation against actual audience, we'd certainly be able to make that judgment. As it stands, it will take Time and the Open Market to find out.
I would love to see publishers out there believe enough in the value of their own audiences to start selling it to media buyers instead of the vague and indeterminate advertising event known as the impression.