Did Data Let Us Down?
The United States later this month will install Donald J. Trump as its 45th president. Regardless of how you feel about the outcome of November’s election, the one thing on which everyone can agree is that this time around, the political polls fell miserably short. Depending on which polls you read, Hillary Clinton had anywhere from a 70 percent to a 99 percent chance of winning the White House. The fact that the polls were so far off the mark (at least on the state level) led to a lot of anger, and an article in The New York Times even went so far as to declare that “data failed us” in November.
A lot of the anger stems from the fact that people in almost every walk of life put a lot of faith in data today. They have become obsessed with data. They expect that with the amount of data available, they should be able to accurately predict the outcome of just about anything—from a presidential election to a football game, from a sales meeting to a TV ad campaign.
But it’s important to keep things in perspective. In our haste, we just might have lost sight of data science’s inherent limitations. While data science has tremendous potential to uncover business insights never realized before, it is still a very new field of endeavor with very few skilled practitioners. In addition, the technology that makes data science possible still comes with very significant trade-offs, least of which is an inability to pick up on the context and subtle nuances contained in the information available. Data science, as we know it today, is still very much based on what’s there in black and white, often impervious to the hints of gray that lie under the surface.
But that is starting to change, as Associate Editor Oren Smilansky points out in this month’s cover story, “10 CRM Trends to Watch in 2017.” One of the leading trends he uncovered was the growing use of predictive analytics, powered by all sorts of advanced technologies—such as artificial intelligence, machine learning, deep neural networking, and cognitive computing—to help businesses determine not just what people are doing but why.
With the advancements in sight, companies will be better able to make detailed predictions about human behavior—from how we’ll vote to what we’ll buy or which type of marketing email we’ll open. Ultimately, companies will be able to use data science and analytics to prove or disprove assumptions about the future and then to craft prescriptive actions likely to result in optimum business results. This goes far beyond the basic insight available today, providing foresight and a path to act on it.
Still, while data can and will be extremely helpful, we can never lose sight of the fact that data science is only a tool. As the election proved, we can’t be too quick to make decisions based on data alone. Forecasting, in politics or in business, is still very difficult, and the results can never come with a 100 percent guarantee.
And while your eyes invariably turn to the events in Washington later this month, why not make plans to visit the nation’s capital yourself this spring for the upcoming CRM Evolution conference? Now in its 12th year, CRM Evolution 2017 will be held at the Washington Marriott Wardman Park, April 24–26. The conference program is shaping up nicely, with some of the brightest and most influential people in the industry scheduled to speak. We’d love for you to be there as well.
Leonard Klie is the editor of CRM magazine. He can be reached at email@example.com.