Google Trends, Robert Jarvik and Pharma Marketing

By March 13, 2009Uncategorized

One of Ryan TrueHealth's constant themes is patient-centered marketing. Distilled down to its essence, this concept means that pharma marketing has to be based on a deep knowledge of the patient, rather than an explication of the features or capabilities of the drug. This strategy is absolutely common practice in other consumer-driven industries (if you've ever bought a car and then been surveyed about your experience the next day, you know what I mean) but pharma, as in so many things, is different.

It's simply not as easy in pharma to understand what your patients (and potential patients) are thinking, feeling and doing. First, there are a lot of factors that affect how a patient thinks or feels about his condition and the drugs for it. Some of these are connected to the drug, and some are not, but may be attributed to the drug anyway. Second, privacy is an issue — you can't simply start surveying, say, MS patients at random. And third, because gathering information is often such slow work, it's not easy to get a meaningful sample, and thus, meaningful data.

Enter Google. In a fascinating item in today's Wall Street Journal Health blog, there's a post about the use of Google search data, particularly Google Trends, to track potential outbreaks of disease. Specifically, in Canada, there was a correlation between Google searches for the term "listeria" and the reporting of an outbreak of the food-borne illness that eventually resulted in 20 fatalities.

The same technique might be applied to evaluating the effect of events on brands. Admittedly, it's a somewhat crude measure, but it's also one done across a vehicle that processes over a thousand searches per second, so if there's data to be mined, there's no better data set than Google.

As an example, let's consider the controversy last year over the use of Dr. Robert Jarvik as a spokesman for Lipitor. We blogged about this in the past — essentially, in case you've been living in the woods for the last year or so, Jarvik was hired, then fired, as Lipitor's televison spokesman when it was learned that he wasn't, for example, a cardiologist who actually saw patients. It was something of a scandal, and if you take a look at the relevant Google Trends report, the effect it had on search patterns was clear. There is a huge search spike the day Pfizer announced Jarvik had been terminated.

However, if you look at the trend results for "Lipitor" for the same time period, you see something different. There is a minor increase in searches for that term early in 2008, when the Jarvik story broke, but it's nowhere near as significant, it abates quickly, and search levels go back to their original level. Initially, at least, this seems to suggest that although there was interest in the Jarvik story, it wasn't generally connected with Lipitor itself.

Essentially, this is a kind of generalized version of the ComScore method of tracking audience behavior. ComScore's system relies of an enormous compensated panel that allows their Internet usage to be monitored, then extrapolated into conclusions about the behavior of the Internet audience as a whole. For example, if 7% of the ComScore panel visits eBay once a month, it's intepreted to mean that 7% of the Internet audience as a whole does.

ComScore has a much small sample, but much more rigorous tracking of behavior. Google's tracking, at least as far as we know, is much less sophisticated, but on the other hand, it's the result of the genuine searching behavior of the audience itself. Since the audience, and Google, are only getting larger, it will be interesting to see where all this goes, and what impact it has on how pharma companies study, and respond to, their patients' needs.

Leave a Reply