Doing the Math on CPMs

There’s been lots of talk about CPMs lately, some of it well-informed, some of it a bit naïve. Most of this chatter concerns two good research reports, which I’ll comment on here. But first, it’s important to clarify two important things about CPMs that often get lost. First: CPMs are a measure of audience quality. Second, CPMs are a matter of yield management. That said, CPMs are often not easily compared quarter to quarter or even year to year -- so hold off on the pizza party, unless your CPMs are up and you’ve satisfied your goals with regards to both audience and revenue per page.

Let’s take first things first. Last week’s PubMatic survey shows that the average CPM of its non-guaranteed inventory rose 111 percent from Q4 2008 to 2009. Bravo! Even if that CPM had previously fallen to under $1 (which it may well have) anything that shows an increase in the inventory normally accessible to third party networks, or inventory that may have gone unsold by premium vendors is most likely a good thing. But the audience is still key. CPMs generally increase because the value of the advertising is increasing to advertisers. It’s saying that the pages may be fewer, but they’re more valuable because the audience available to read them is worth more. Advertisers pay for the quality of audience. CPMs reflect that quality.

Now let’s talk about yield management. Yield management involves adhering to a strict discipline of serving the advertisement which will increase the network or publisher’s revenue per page and RPM. There are, now, hundreds of companies and scores of technologies that help publishers measure RPMs, in real time, and ad serve the optimal ad each time there is an ad call. Many publishers have, only recently, begun to yield manage their remnant inventory, in a precise manner, through technology. Many publishers have introduced this stepped-up process immediately after their brand deals are served. Most often they are served in priority order tied to CPM value. Many publishers are also using dynamic rate cards on brand deals in order to eek every bit of value out of their ad inventory. CPMs reflect this management.

What encourages me is that everyone today seems to be focused on CPMs. When I see PubMatic’s numbers, and I see the recent JP Morgan numbers that predict a 10 percent CPM increase for display ads this year, I see three things at play. One: there are less remnant pages available. Two: the Internet audience is increasing in value. Three: effective yield management is at work. If your company is tracking well against these three metrics, don’t worry about the CPMs. They will take care of themselves.

On a side note… I’ve heard it said that some premium publishers would like to get away from CPM pricing models and move more toward project-based pricing. That may be more retro than it sounds. When I worked at Yahoo several years ago, we dabbled with this model with home page takeovers and guaranteed-rotation, performance deals. I’m sure you remember the “slotting” fee! Project pricing does have its place in the market and we, at FOX, sometimes use this pricing model in our business on popular, niche sites that advertisers like to road block, like Smartsource.com. Project based pricing and CPMs can’t really live in the same house – unless, of course, you have a team of analysts feeding you real time pricing data.