Dirty Little Secrets

It's no secret that I don't come from an advertising background. I do come from a finance background but that still doesn't explain my shock at what I've discovered in online advertising: nothing adds up. Yes I said it. And it's not a new subject apparently, just one many people don't talk about. Such as publisher reach numbers and ad network reach numbers. None of them make sense. Why? Because server logs don't equal Comscore or Quantcast or other third party methods. And it's no wonder when you get behind the methods. Yet advertisers and agencies, with few options, are forced to rely on these third party stamps of approval. Therefore I have to pay tens of thousands of dollars just for a little dose of legitimacy - but it's a farce - anybody can pay and the numbers mean nothing!

I mean does anybody find the fact that a major third party traffic analysis firm was grossly wrong about a large public company's traffic numbers even slightly incredible?

Even the fact that different ad servers consistently come up with different numbers is enough to drive a person insane. Apparently each ad server has a different definition of "delivered."

And if traffic numbers and ad serving impression numbers weren't enough, there is a whole host of plain ol' bad math. For example, the number of Hispanics online - how is that measured exactly?

Have you ever looked at the sample size of some of these research reports? Where n=800. 800 people? This is a sufficient sample size to extrapolate to the whole of 43 million Hispanics?

But why stop at traffic? How crazy is it that publishers don't know nor do they have a good way of knowing who it is exactly that makes up their audience? People always ask me how publishers can know this information and I reply: four ways: they have a registration path where they collect the information, they've surveyed their audience and extrapolated to all visitors, a third party analysis company has surveyed their site or used the panel system to extrapolate to the publisher's audience attributes or the publisher makes it up. That's it.

The reality or dirty little secret is that many publishers don't know for sure. And this information seems to have eluded most agencies and advertisers. I can't tell you how many RFPs we see where the agency/advertiser is looking for brown-haired, males, aged 18-25, interested in autos, in South Dakota, with only one brother. The specificity can border on the absurd and in most cases is completely unaddressable - exactly how are you supposed to tell if your audience has brown hair? Well the horrible secret is that it's just as hard to determine their age.

At Consorte - we ask, and we ask, and we ask. But we're only as good as the information given to us. And that's the horrible state of metrics and data in online advertising. Which considering the source is utterly disappointing. One of the beauties of online is its supposed measurability!

So, what's the alternative? Media Strategies Editor Jim Meskauskas in a recent posting suggested that advertisers and agencies simply refuse to use the numbers. I'm not sure that's realistic. What do you think?