Monday, December 15, 2008

In Defense of Good Data

Last week in Marketing Daily, Adrian Chedore, CEO of Synovate, made an argument for the more robust use of big data sets along with smaller, in-depth studies of consumer behavior. One of the challenges, he suggested, is being able to make use of the massive amounts of data that we are now able to collect. Putting aside my objection to reducing the full measurable reality of people’s lives via their consumption habits, it’s important to recognize that information is useless without two things: reliability and interpretation.

Reliability means we know the information is good. I know this is a standard social research rant, but every day there’s new reports touting data that proves what people want, who they are, what they buy, based on sample sizes and questionably designed survey techniques. In a frantic attempt to name and claim segmented markets, research firms will prematurely christen fragments of the population with ridiculous nomiclatures. I spend a fair amount of my research and reading time figuring out the design and sampling procedures of studies. Even the good ones tend to generalize too far out from their data. Intepretation means what ideas are being used to make sense of the data. I'm also constantly digging to make sure that the studies cited really do measure what they purport to measure. In the desire to say something new and useful, the data often gets left behind. For a good explanation of how to decide what statistics are good or bad, here's a recent broadcast from noted sociologist Joel Best on Kojo Nnamdi's radio show.

Given that, I’m in agreement with Mr Chedore, but I want to put a plug in for the most important source of data we have: one that is not generated by commercial entities, but by the government. It’s the Census. Recently the New York Times included an editorial supporting the new administration’s concerns about the 2010 census, specifically about finding a director with proven experience. The last eight years have seen reduced funding and administrative upheavals at what was once a very reliable agency.

Indeed, Chedore insists that market research's strength lies in its foundation in academic discipline and commitment to sound, reliable data. He pushes for more coordinated use of data and international comparisons – something that might be more effectively handled with government-generated data rather than across different companies. Case in point, Chedore laments the highly fragmented nature of current market research efforts. Using the Census might be an important corrective.

The census is extremely important to all of us who are interested in the demographic makeup of this country. It’s also necessary as an unbiased benchmark against which we can compare other kinds of data. In January, John Tizzi reminded the business world that the census is the best free market data around, another smart tip in lean times. The public accessibility of the data is one of the most remarkable things about it. (Tizzi’s other suggestions for “marketing on the cheap” are actually some very sound, basic research principles that companies seem to be forgetting in their constant search for something new).

So it may be smart for companies to put in a good word for the new administration's efforts to revitalize the US Census. After all, there's nothing like good data.

Map above is from Google Earth's census mapping program: it is a map of the counties in the United States colorized by median age. Lighter colors are older.

No comments: