What is Predictive Analytics? An Introduction

Predicting outcomes - will it rain?

Probably the most common way to introduce the concept of predictive analytics is to use the weather report - and it's a great analogy for the layman: based on cloud patterns, atmospheric readings, wind direction and a long list of other meteorological data, we are provided with a short and long-term forecast, and it has become increasingly reliable due to scientific advancements - especially when I think about the accuracy of weather predictions in my father's day, a retired air force navigator who used to say that a 50% chance of precipitation meant that half of the folks at the weather department think it will rain.

But perhaps a deeper look at predictive analytics is in order, even as we examine its influence on outcomes in popular areas like sports and entertainment.

In his bestselling book, Moneyball, Michael Lewis demonstrates the power of predictive analytics as he tells the story of the Oakland A’s manager, Billy Beane, who was constrained with the lowest team budget in Major League Baseball. Beane used predictive analytics techniques to orchestrate a dramatic turnaround in his team’s performance.

NUP_152770_0994.JPG NUP_152770_0994.JPG

NBC's The Voice brings a new concept to talent competitions, by using blind auditions to remove bias from predicting outcomes, allowing the judges to concentrate on a performer's voice without any visual distractions. This example outlines a specific area of predictive analytics that deals with removing any subjective influence from the results of a data set analysis. And it's a perfect example where intuition can sometimes interfere with a logical choice. Note that The Voice is not the unique purveyor of blind auditions in the music business: since adopting the same concept for their grueling audition process, symphony orchestras around the world have seen a 50% increase in the advancement of women in their ranks. 

With these examples in mind, let's move on to the next step, finding a solution to your issues with a specialized approach.

How is predictive analytics going to help me solve call center challenges?

To answer that question, it's vital that what appears to be an obvious task is completed ahead of time. List the challenges in your contact center that keep you up at night. Based on previously gathered data from surveys and conversations with peer groups, allow us to start a sample list:

  1. Reduce churn/attrition, especially with talented, high-performing staff.
  2. Increase performance and compliance among contact center agents.
  3. Forecast and plan staff schedules with better accuracy.

When looking for a way to increase the performance of your call center in these areas, consider the data that needs to be collected. Is the solution considering past, historical data? Are there correlations between compliance and performance? What about the human factor, such as job satisfaction? How is it being considered in the mix?

Our next post on this topic will examine the concepts of desired outcomes, inputs, validation, and deployment - all of which are steps that are required for a data analytics model to be successful.