When web ad measurement fails, build a ruler

We've been critical about broadcast, outdoor and print media measurement systems lately ... but at least there is one ruling body in each of those media that tries to establish standards. A new report says the lack of consistency in online measurement may be scaring some advertisers away from internet advertising, especially as web media continues to evolve into video, widgets, and social media apps. This is ironic, because the web is awash in data ... but often not all of the numbers add up.

For example, one standard of web audience is "unique visitors." But if a user deletes the cookies on his computer, he comes back as a new unique visitor each time ... so John Doe, battling viruses on his Windows PC, may get counted over and over again inflating a web site's traffic report. "Uniques" also do not account for the same viewer logging in from different computers.

This is more than a nit: unique visitors are the baseline of a web site's audience, the equivalent of reach in traditional media planning. If reach is off, what else could be?

Randall Rothenberg, CEO of the Interactive Advertising Bureau, put it: "Marketers want to know, If I take $10 out of TV and put it into online, am I getting $10-plus back?"

The cold hard truth is the advertising industry has been riddled with inflated metrics for years, and as the data systems get better (driven by the attempt to keep up with the flow of data from the web), a lot of old-school Emperors are going to realize they are wearing no clothes. Media can work, but only if you base your media plans on reality. If the reality is teenagers are leaving radio to listen to MP3s, let's count it, and deal with it, not hide facts. If the reality is newspaper circulations are down, let's not make up bogus "readership" claims based on passalongs, because that does an advertiser little good.

The only real solution we see is for marketers to bypass the competing claims by setting up measurement of inbound responses. If a marketer spends $10,000 on one outlet, and can track with accuracy that 200 responses came in, she will know the cost per response was $50 in that particular media. If another media option drives a response at $25, then that's a better deal. As media options continue to proliferate, each channel will inflate and defend its own metrics, and silly claims will continue to rise. If you want the truth about advertising, measure it yourself.