By Paul Stockford, Saddletree Research | Jan 17, 2017 | Comments 0
“Predictive analytics isn’t yet a mass-market solution, but it’s only a matter of time before the mass market realizes the power of predictive analytics and that changes.”
One of my favorite things about the beginning of a new year is reading all the predictions that the soothsayers, psychics, wizards and other prognosticators foresee for the year ahead. It’s kind of like reading a horoscope for an entire year and for the whole world. In the deep recesses of your rational mind, you know most of this stuff doesn’t really end up happening, but it’s still fun to read about.
History tells us that even the experts don’t always get their predictions right. For example, in 1977 Ken Olson, the founder of Digital Equipment Corporation (DEC) predicted that, “There is no reason anyone would want a computer in their home.” Apparently, he didn’t foresee the inevitable invention of Netflix.
Still, I admire the psychics who put it all out there and make the bold predictions, especially the ones that take themselves so seriously. At the end of 2015 Tana Hoy, who bills himself as “The World’s Foremost Psychic,” published his list of predictions for 2016. Let’s review, shall we? And I quote:
“Prediction is very difficult, especially about the future.”
—Niels Bohr, Physicist
“Hillary Clinton will win the upcoming election and become America’s first female president.”
“The last living Bee Gees, Barry Gibb, will come down with some serious, if not profes life-threatening, health issues.”
Actually, Barry did 78 concert dates in the U.S. alone during 2016. Not bad for a seriously sick guy.
“Many parts of Arizona are going to face record-breaking temperatures during the upcoming summer.”
I’ve lived in Arizona for 21 years. We break temperature records in one way or another EVERY summer.
“A Republican Senator will be caught taking bribes and payoffs, and will end up spending some time in federal prison.”
Predicting that there will be crooked politicians is like predicting that the sun will rise in the east tomorrow.
Let me add a provocative prediction of my own to Tana’s list: In 2017, customers will call a contact center. I’ll even make my own bold industry prediction: Of all the technology solutions that will affect the contact center industry in 2017, the one that will have the greatest impact is predictive analytics.
As a solution, predictive analytics borrows a variety of statistical techniques from modeling, machine learning and data mining in order to analyze facts and make predictions about unknown future events. For example, if you’ve ever financed a major purchase such as a home or a car, predictive analytics have been used on you via your credit score. A predictive model looked at your credit history and other personal data to determine the likelihood of your making credit payments on time.
In the contact center, predictive analytics are employed to predict customer behaviors or business outcomes based upon the analysis of mostly unstructured data gathered during a customer interaction. Predictive models are created based upon the results of mining unstructured data and testing the different categories of the data that are uncovered in order to determine which data points are important and which are not relevant to the outcome the predictive model is testing for.
To learn more about how predictive analytics functions in the contact center I turned to Larry Skowronek at Nexidia, a NICE Company. I’ve known Larry for more than 15 years and not only is he a longtime business colleague, he is an authority on analytics of all descriptions in the contact center. According to Larry, “10 years ago, analytics wasn’t even a buzz word, much less predictive analytics. Today analytics are everywhere as businesses experience challenges around maintaining compliance, reducing churn, managing costs, improving customer satisfaction and increasing sales effectiveness. At the same time, as an analytics company, we also need to rise to the challenge. We are evolving to the point where we no longer program computers to do something, we program them to learn how to do something.”
Fortunately for customer service professionals there are people like Larry and his colleagues at Nexidia to do the heavy lifting when it comes to predictive analytics. They develop the predictive models based upon specific customer data which in turn provides contact center managers with customer insights that may otherwise go undiscovered.
“Customer churn is something that we address quite a bit with predictive analytics,” Skowronek explained. “Customers cancel services for a variety of reasons. In order to accurately predict both the potential for churn and the reason for cancellation, we start with capturing and organizing 100% of the interactions across all channels. We then analyze each individual transaction for any number of scenarios, phrases, sentiment… etc. The more data captured and analyzed, the smarter the system becomes at identifying at risk customers allowing companies to respond sooner with the right answer. Before it’s too late.”
Given the state-of-the-art today, predictive analytics is on the cusp of radically changing the way both customer service functions and marketing departments address customer needs and concerns. Combining the typically unstructured data of the contact center with the typically structured data of the marketing department results in unprecedented insight into customer behavior and unmatched accuracy in predicting customer behavior. It’s going to change the game.
Like most of its cutting-edge predecessors—speech analytics, for example—predictive analytics isn’t yet a mass-market solution, but it’s only a matter of time before the mass market realizes the power of predictive analytics and that changes. Given the potential impact of predictive analytics in customer experience applications, I predict 2017 will be a big year for predictive analytics in the contact center.
Paul Stockford is Chief Analyst at Saddletree Research, which specializes in contact centers & customer service.
– Reprinted with permission from Contact Center Pipeline, http://www.contactcenterpipeline.com