For years there’s been an ongoing debate regarding the veracity of research in the advertising world. On one side of the argument are the research proponents who believe that consumer attitudes and purchasing preferences can be accurately determined through research.
On the other hand are the anti-research critics who believe that the creative process, and our response to it, is too emotional, esoteric, and erratic for us to understand it ourselves, let alone be able to explain it to anyone else. As the anti-research group says, market testing is like dissecting a frog. It’s bloody, it’s messy, and it’s not real good for the frog.
But now there’s a new modality that might change all that.
Brought to you by the ambitious folks at Google, Google Correlate allows researchers to use seemingly disparate consumer decisions to generate statistically relevant data. And it’s available to anyone who logs on to the site.
According to the Google Correlate introduction, the service “finds search patterns which correspond with real-world trends.” Using the free service, researchers (and you) can look at a number of ways data relates to other data. Specifically, you can determine:
1. How search items vary in popularity over time;
2. Which items have a similar pattern of activity;
3. An item’s pattern of activity across states; and
4. How specific searches correlate state-to-state.
What does this mean to researchers (and you)? Quite simply, we now have access to information that can be used to determine consumer activity based not on buyers’ responses to theoretical queries or imagined courses of action but based on their actual and documented behavior.
What can you actually do with Google Correlate? The example Google uses shows how they were able to track the pattern of the influenza breakout as it spread by tracking the volume and types of questions asked about the flu across the country in real time.
NPR reporters Linda Wertheimer and Shankar Vedantam report that by comparing the volume of searches done for liberal news commentators with the searches those same questioners made about food, University of North Carolina researcher Phil Cohen determined that liberal Democrats are interested in “arugula pasta, beets nutrition, beets urine, fake meat, fennel salad, firm tofu, a variety of vegetarian cooking, (and) vegetarian recipes. Something like a Republican stereotype of what a liberal food diet might be.”
Vedantam goes on to postulate, “…when we think about our political orientations, we tend to think that our ideologies determine whether we’re Democrats or Republicans. But…what this research is at least hinting at, is the possibility that our political orientations are really a matter of our identities, are a matter of our cultures. And so if you’re somebody who’s a vegetarian, who likes beet salad, it’s very unlikely that you’re going to be a Republican.”
Whoa! While this assumption does have the comfort of stereotyping to back it up, it does start to move users of Google Correlate back into the argument of whether or not marketing research is accurate and usable. After all, as Google — and decades of scientists and statisticians — have pointed out, correlation does not imply causation. Nor does correlation prove causation. So just determining that people read more Stieg Larsson novels at the same time that the obesity rate increased, for example, does not suggest that Swedish murder mysteries make you fat.
Still, by replacing human inconsistencies with measurable statistics, Google has given us all a formidable and heretofore unavailable research tool. And did I mention that it’s free?