So Google thinks they can measure the ROI of advertising in offline media. Good luck! For years, Corporate America has been putting the screws to marketing departments to find a way to measure the ROI of advertising, but until recently the results have been mostly smoke and mirrors.

One of the (many) empty promises of the 90’s Internet bubble was that advertising would finally be measurable and accountable — but clickthrough rates just didn’t cut it, and everyone ran for the exits.

Then came Google AdWords, with 1) text ad formats that people actually clicked on and 2) the ability to manage the cost of ad campaigns against actual returns — savvy search marketers could track the conversion of clicks to sales. Suddenly, advertising was transformed from a cost center into a profit center.

So I guess you can’t blame Google for wanting to bring the miracle of science to the art of offline advertising. Media Daily News got the scoop from Google’s Director of Advertising Strategy Patrick Keane on the plan to “provide a new, empirical way of proving the ROI of traditional media advertising deals”:

…the plan relies on techniques already being utilized by some advertisers who use an online component of their media strategies to measure the effects of traditional media in their mix.

One straightforward strategy, he said, would be driving readers or listeners back to the online realm and measuring them there.

“The smart advertisers have been coming up with linked campaigns for a while,” Keane noted. “They’re no longer conceiving of advertising campaigns that are limited to the various silos–just print, just radio, or just Internet, and so on.”

As an example, Keane noted how marketers might include a unique URLs in text ads, allowing advertisers to measure by site visits the number of visitors who interacted with the text ad. He suggested advertisers could cast an even wider net by including an old direct response technique–the unique 1-800 number–in print campaigns and radio advertising. As aficionados of late-night TV know, 1-800 numbers are old hat; in this system, the studio selling albums, for example, pays out to the broadcaster based on the number of phone inquiries they receive.

An 800 number on a print ad — now there’s a innovation! Compare: Clicking on a URL in an online text ad vs. typing in a coded URL lifted from a print ad — the latter is going to tell you what exactly about the overall impact of the ad? Are most people going to take the time to type in a URL? This type of “direct response” mechanism has always been a less-than-user-friendly way to respond (since it was conceived of ages ago using 800 numbers), but thanks to Google and the link-drenched online world, most people have been trained to expect one-click gratification.

Memo to Google: Offline advertising — with the exception of infomercials — is a crumby direct response vehicle. Advertisers made due when that was all they had. But Google is delusional if they think these crude direct response mechanisms are going to create an ROI measurement capability for their Print Ad program that is even a pale shadow of what they can do online.

But it gets better. If direct response doesn’t work, there’s also “econometric modeling”:

Another approach might include econometric modeling, which uses sophisticated statistical analyses to determine the effect offline advertising has on driving consumers to online activity and vice versa.

John Nardone, client relationship director at marketing mix modeling firm Marketing Management Analytics (MMA), a unit of Carat parent Aegis Group, says the firm already is doing similar work for Google rival Yahoo.

“By looking at GRPs for television and TRPs for print advertising over time we can determine how those are driving online traffic and sales.” Nardone declined to give details of his work for Yahoo, but said Google might be looking at a similar approach.

Marketing mix modeling (i.e. econometric modeling for marketers) is the ultimate black box — feed in lots of data about advertising and “other” factors that drive sales, wave the magic wand, and presto — advertising ROI.

I won’t bore you with a detailed debunking of marketing mix modeling. Here’s the problem in a nutshell — garbage in, garbage out.

If you had “perfect data” on advertising activity, other marketing activity, and all other factors that influence sales, e.g. seasonality, world events, day of the week (it’s endless), then you could use these statistical modeling techniques to back into the ROI of advertising. But, unfortunately, the data sucks.

Here’s the definition Gross Rating Points (GRPs) from Nielsen (which holds the entire TV ad industry under the thumb of its highly suspect data):

The sum of all rating points achieved for a particular period of time and/or schedule of commercials or spots (called Gross Impressions when expressed as the sum of all exposures to a given time period or spot schedule).

Got that? Bet you wish you were back in the land of click and buy.

To be fair, I’ll give Google some credit on the digital radio front:

Keane pointed to dMarc’s much-touted ability to verify ad play “down to the second” through its digital tracking system–but was close-lipped when asked whether the rollout of digital radio, in both terrestrial and satellite forms, might hold promise for tracking consumer response as well.

Nonetheless, digital radio technology firm HD iBiquity’s CEO, Bob Struble, has spoken seriously of including “buy buttons” in digital radio sets, allowing advertisers to establish causal links between ads and purchases. DMarc CEO Chad Steelberg confirmed in January that dMarc is an active partner of HD iBiquity, and is following the rollout of digital radio closely.

But then the Media Daily News article ends with reference to “the emergence of interactive TV–capable of both targeted advertising and tracking consumer response–seems like a promising area for Google to implement a better ROI model.”

Interactive TV — no there’s a panacea. Seems like a “promising” windmill to tilt at.