Driving home the other night, I heard a political analyst on one of the news radio stations explaining how the Republicans were moving some big money out of some US Senate contests and into others, while Democrats were doing the same. Some of the races they were pulling money out of were considered "safe wins" with only a week of campaigning left. Others were forecast to be lost causes where more TV ads weren't likely to turn the tide.
I found myself marveling for a moment at the confidence they must place in their polling techniques to develop such levels of certainty about outcomes before the first votes are even cast. But when I really thought about it, their certainty isn't wizardy, it's just the science of commitment.
Political organizations have been doing polls for well over a century now. And to their credit, they've done two things right along the way: always updating their technology; and cataloging their knowledge. Think about it for a minute. They're masters are adapting to the latest technologies - from telephones to televisions to internet samples. They learn fast through repetition and quickly iterate to a sound understanding of how to read the bias and variance in each tool.
But more importantly, they're great students of history. They keep poll results going back decades and constantly update the analysis of the historical perspective so they can find any possible parallels between today's races and victories or losses they've suffered in the past. Their knowledgebase is both expansive and accessible. And next to a charismatic, articulate candidate, it's their biggest asset.
So what can we learn from the pollsters about enhancing marketing measurement and predictive ability? Well for starters, we should learn that you CAN get very confident in the predicitve value of survey-based research if you have the right methodologies and stick with them long enough to understand the underlying variances and biases in the answers. Regularly comparing reported intentions to actual behavior patterns, even amongst a broad and anonymous population, can identify predictable levels of difference between the two, in effect providing a perfectly predictive mechanism for anticipating future behavior. That sense of history provides a significant competitive advantage that simply cannot be copied.
Second, we can probably learn that even in these days of amazing analytical tools, survey research is critical to our understanding of what works and why. This suggests that we should reconsider the role of the chief researcher in our marketing organizations and ask a few simple questions, like: what questions do we need this person to help us answer? What is the economic value of higher levels of certainty in the answers we get? What resources do we need to give them to permit them to achieve their mission? Do we have the right person/people in charge of this critical function - not only technically competent, but strategically saavy and organizationally influential? And are we willing to commit to consistently improving our research month by month to get closer to understanding and prediciting things we can't see in consumer/customer behavior?
It stands to reason that if marketers are as committed to learning from survey research as the pollsters have been, we too might have the informed confidence to move large sums of money from losing efforts to more promising opportunities in time to win.
What do you think? Is research able to play a role like this in your marketing organization?