But this approach does not give advertisers a transparent view of the value of their digital advertising, nor is it an ideal way for publishers to price inventory.
We believe that impressions are a largely meaningless metric for advertisers. They do not tell advertisers anything about the benchmarks they should really care about: How many unique people are seeing the advert and who are they? How often are they exposed to the advert? And what share of voice is the advertiser getting from a campaign?
Another problem with impressions lies in the fact that they can easily be manipulated through techniques aimed at multiplying page-views without improving a publisher's audience or the audience's visit duration. Impressions are not proportional to reach and frequency, measures that should be familiar to advertisers.
Multiple advertisers will often be jostling for a user's attention on a web page, so should you be paying the same CPM rate when you're one of four advertisers on a page as you would if you were the only one?
Against this backdrop, digital publishers should start moving towards a model where they report more accurately on unique users and provide detailed information about the true reach and frequency numbers they offer advertisers.
Online media planners and brands, meanwhile, should apply the same fundamentals to digital as they do to other channels. They should look carefully at reach, frequency and campaign length when they purchase digital advertising inventory.
The concept of impressions still has its place in network buys and other low-value placements, but we believe that advertisers should start to move away from it in their more strategic campaigns. They should be looking at how they craft content and environment that are conducive to the conversations they want to have with their audiences, and then look at its value using many of the same metrics they'd apply in traditional media.
