The Enduring CPM and its Discontents

The Internet is a breeding ground for unlimited punditry, and the pundits are almost always wrong – as in the area of Internet advertising in which self-proclaimed seers declare the demise of the CPM, the fundamental unit of measurement in Internet Advertising.

The world has changed, they say, and advertisers/marketers want action and engagement, not just impressions. In this “theory,” the CPM is dead, but the CPC, CPL, and CPA are alive and kicking.

For anyone, however, who has lived a day in the life, all of these fancy contrivances are essentially equivalent. Any CPX is just CPM with a different risk calculus and a veneer of legitimacy.

Yes, Marketers want action and engagement but ultimately the method of calculation that determines the cost of the CPA or Cost-Per-Engagement (CPE) is based on the enduring CPM.
Since this alphabet soup of acronyms is confusing, here is a brief primer on advertising math:
CPM= Cost per Mille or Cost per Thousand Impressions
CPC=Cost per Click
CPA= Cost per Action
CPL=Cost per Lead
CPE= Cost per Engagement

The Display Ad is the simplest form of Internet Advertising. Here, Publishers charge the Advertiser a certain amount of money per thousand impressions, i.e. they charge the Advertisers a certain CPM. It’s in the Publishers’ interest to have a high CPM so they generate more revenue. Theoretically, it’s in the Advertisers’ interest to have a low CPM but in reality if they want to advertise to the “right” audience who is “qualified,” the CPM will be accordingly higher and rightly so since every impression has more chance of being concluded in a “sale” or some other desired action.

The controversy around CPM is not about a disagreement on the mathematics but about the perceived desirability (or lack thereof) of an impression versus a “real action” that shows the consumer is engaged. Enter the CPXs (as above) that each measure a tangible outcome. In the case of CPC is the clicks being measured in the cast of CPA it’s a predefined action (like downloading a white paper, creating a profile, etc) being measured.

This controversy has correctly lead publishers to attempt to create more engaged audiences, ones that will take further action when being exposed to an impression. Thus, they can quote Advertisers not just a CPM but a CPX of some sort, since they are confident that their audiences will either click (as in CPC) or take some action (as in CPA.)

But the fact remains that they still have to quote a number, which they have to calculate. And in every case I have seen, they calculate the CPX based on the estimated percentage of people who will go from Impression to Click to Action or Impression to Action. Thus, ultimately, all CPX is really based on CPM.

CPX = F(CPM) so to declare the CPM dead is disingenuous.

Now, back to the notion that Publishers have interest in keeping the CPM high:
What needs to be understood, by far the most important thing, is that publishers are doing themselves a huge disservice as regards the “C” in the CPX construct. The strategy of most publishers hurts both their advertisers and themselves. The problem is one of simple statistics, the most often misunderstood discipline in our innumerate society.

Simply put, publishers (in most cases, there are exceptions) continue to want to “grow” their uniques and their databases. For a publisher with a small audience footprint, this is fine. The moment, however, one passes a certain threshold, the audiences and unique individuals the publisher bandies about and “sells” to the advertiser start to approximate– in all relevant ways– the entire population (either of a particularly defined audience or of the population in general.) When this happens, the value of a particular publisher’s audience ceases to be relevant as a differentiator.

A telling conversation I had with network exec recently brought the point to bear. He claimed that his network had 100M+ uniques. He claimed simultaneously that “his” audience was far more qualified than that of his competitors. How a swath of 100M people (and that too mostly in the US) can be any different than the population in general in terms of buying patterns, propensity to act on an offer, or any other measure of “value” defies math. It’s not possible. Well, okay, it’s… possible. If for instance all of the 100M are men (how would they really know?) then one can argue that the value-making patterns of this group might be slightly different than that of the general population. But, the population’s size diminishes the differential value of their [male] audience vis-à-vis the male audience of a competitor. If Facebook claims its 400M+ membership is more qualified or more densely valuable than the population at large, you should smell a fish.

Statistics don’t lie. Only people do.

So while the CPM is not dead, most publishers are slowly killing the C. In attempting to relentlessly expand their audiences, they hurt their own businesses and simultaneously provide watered-down coverage for their advertisers.

Thus, the monochromatic desire to grow on the elementary dimension of “audience size” or “uniques” is perilous-it leads to a process of equalization which makes all publishers’ (or networks’) audiences essentially equivalent.

What really matters is the degree to which a publisher can captivate its audience and therefore capture an increasingly larger portion of the audiences’ “economic flow.” This is the hard part and, frankly, the part most interesting to a good marketer or advertiser. It’s about depth not breadth, about quality not size.

The greatly exaggerated reports of the CPM’s demise can be dismissed; the diagnosis of the C’s degeneration, however, should not be. The cancer, if left unchecked, can spread quickly.