What Data Can’t Do
The title for this post quotes one of today’s New York Times editorials. It perfectly captures something essential about our work that we need to trumpet as loudly as we evangelize the power and benefits of research. Data can’t do everything. In fact, the bigger the data, the less it can do relative to the sheer volume of data being generated.
Summarizing (and quoting) David Brooks’ column, here are five important limitations of data:
1. Data struggles with the social. It is pretty tough to measure and analyze the emotional and social substance of what goes on among people beyond the number and type of interactions they have.
2. Data struggles with context. In fact, at Versta we argue that data have no meaning outside of the questions that need to be answered—questions that can only be determined by a human brain in a social context looking at the data from the outside.
3. Data creates bigger haystacks. As data grows exponentially in volume, the 5% of our statistical estimates and correlations that happen by random chance also grow exponentially in volume. “Falsity grows . . . the more data we collect.”
4. Big data has trouble with big problems. When it comes to solving the biggest problems a business might face—lack of leadership, an unmotivated workforce, insurmountable debt, an out dated offering, etc.—data can point to problems, but rarely does it offer up the solution.
5. Data obscures values. Numbers lend an air of objectivity and hard reality to issues that are inherently subjective, emotional, and laden with values. Data are always constructed, structured, and interpreted “according to somebody’s predispositions and values.” There is no such thing as “raw data.”
True, data is powerful, and discovering new insights that help our clients make decisions is a thrill that keeps us going. But our clients are helped even more when we remind them exactly how far the data can take them, versus where their own judgment and wisdom step in.