Half-life in Politics and Marketing

In his new book "Join the Conversation," Joseph Jaffe explains the faults of the one-to-many model of marketing. The idea goes like this: if you throw a big enough net, you're sure to catch a fish or two. Junk mail is the perfect example. But you can narrow down this model by knowing a little about your audience. Jaffe says:

"One-to-many assumes that it is possible to divide and categorize human beings into generalist buckets, using artificial variables such as age, sex (yes, please), occupation or education, and, to a lesser extent, attitudes, interests, and opinions." (Jaffe, pg. 13)

In a way, this goes against what I have learned in politics. Generalizations, polls, surveys, and focus groups all get you close to an idea of what people want. So much is determined by these methods. It makes sense to do a Get Out The Vote (GOTV) campaign in the district that voted most heavily for your candidate's party in the last election. It makes sense to want to know that when you're talking to a middle-aged, white, suburban mother that you should talk about your candidate's national security stances while you might want to discuss the environment with a local college student (and about their voting eligibility).

The natural result of this obsession is Mark Penn. He is a chief strategist to Hillary Clinton and developed the idea of "soccer moms" for Bill's run in 1996. His new book, "Microtrends: The Small Forces Behind Tomorrow's Big Changes," seems amazingly misguided in this new era. Or perhaps I can just feel this mode of thinking going extinct more and more every day. This NYT article does a good job of breaking down Penn's thesis.

It is against this model that (I believe) Jaffe rebels. In the brave new world of personalization (real personalization), how useful are broad polls or trends? In fact, in an effort to know our audience, don't generalization/trends/polls get us close, but in tandem ensure that we don't reach this last intimacy with our audience?

In science, a half-life is the time it takes for half of the radioactive atoms of a given object to decay. Imagine this is a pie: first we cut it in half, then another half takes it down to a quarter, then another half takes it down to an eighth, continued ad infinitum. But by this method you are guaranteed never to get rid of all the radioactive atoms completely. Despite the small size, there is still half that decay and half that stays.

Likewise, polls and the like get us close to our audience. We do learn important facts that have helped win elections. But this method also guarantees that we will never truly know our audience. By the very manner of data extraction, we are generalizing all the time. A candidate could have a personal connection by talking to Jane Smith. But a candidate can never have a truly personal connection by talking to a "soccer mom," or "NASCAR dad," or any of these other ridiculous groupings of people that make up Penn's microtrends.

It turns out that people are more complex than that. They can't really be grouped this way because we don't think in lock-step. Heck, most of the time we're not even logical. Knowing people is difficult. Perhaps we can learn something at the intersection of science, politics, and marketing.