Shelby Bonnie, founder of CNET, wrote a very interesting article for Tech Crunch in September, arguing that the time has come to for both advertisers and publishers to do away with the CPM (Cost per Thousand Impressions) model. Bonnie offered several reasons as to why the CPM, a pricing model commonly appearing on media plans and often used to decide which properties make the buy, does not work for the buying and selling of online display ad space.  Bonnie correctly identifies several issues while making the case against the CPM, and concludes that the CPM, ultimately, undercuts the opportunity to execute innovative & creative campaigns  However, the leap from outlining the drawbacks of the CPM to it’s termination, is hardly warranted as a solution to these challenges  (or likely to happen).   Where Bonnie’s argument goes wrong lays in the fact that the CPM, which has historically been used as a model for purchasing media long before the days of digital, has never been used as a performance metric by digital marketers.

Why? Data & measurability, two cornerstones of the digital space, are key drivers of advertising effectiveness.  When a campaign is properly managed relative to the end objectives (such as reach, engagement, conversions or revenue) this should mitigate the impact of a single metric, specifically and especially CPM, on campaign success.

For example, he writes:

“If you pay for impressions, you get impressions. Is that, in the end, what marketers really want? How about engagement? How about impact? How about actually selling product? A glut of impressions has helped no one.”

True.  When a user is presented with several ad placements during a single-page view, the advertiser is often paying for 3 or more impressions, which, at best, can only reach, engage or convert one consumer.  If that user views several pages across the same site or network, the ratio of impressions to actual viewers continues to decrease. This potential for duplicitous & duplicative flighting of impressions, however, can easily be considered as the planner makes decisions on how to the structure a media buy.  If not, the publisher may certainly deliver tons of impressions that have little impact on an advertiser’s end objectives.

The major fallacy with this argument is that no consideration is given to campaign management.  What sets digital media apart from its traditional counterparts, is that the planners job is not complete after the creative is trafficked.  If the planner finds that placements on a site or network are receiving “junk” impressions and fail to meet any of the advertisers performance objectives, she should optimize the campaign by moving that investment to placements that perform.

If Bonnie’s contention that the columns for impressions and cost are often viewed as the most important on a performance spreadsheet is true, then there is no reason for publishers not to focus on maximizing the number of impressions sold without regard to quality.    It would also suggest a dilemma much larger than that of the CPM.

Bonnie also argues that the scale and quality of impressions available through digital channels have completely eroded the intrinsic value of the impression.  From the agency and advertiser perspective, however, this and other points outlined by Bonnie do not necessarily represent a reason to discontinue using the CPM model. This is because CPM is a pricing model and not a performance metric.  CPM or eCPM would offer very little insight and value as a lone performance metric.    While there is no one universal metric that does, marketers should look to a combination of engagement, conversion and efficiency metrics to gauge how successful their publisher and network partnerships are at achieving advertising goals.

The industry may not be prepared to deliver a coup de grace to the CPM until a better and convenient replacement is made available.  For now, the CPM model as originally and correctly intended, works just fine.  As a marketing performance metric, it’s worthless.