Back to School With Big Data Analytics

Students are not the only ones who go back to school. We can all come back to learn about ways to direct our efforts more productively. Predictive analytics can show the way. Whether applied to university recruitment or corporate hiring, what big data reveals can show us that our assumptions about what works are leading us in the wrong direction.

Analytics In Action

For those whose business is school, preparing for this season takes planning, and big data analytics can show how to get maximum results. That’s the story of Wichita State University’s strategic planning. A couple of years ago, David Wright, associate vice president for academic data system and strategic planning, sold the Kansas school on using big data analytics to increase efficiency in scholarship spending and recruitment.

"Building a Smarter Campus: How Analytics is Changing the Academic Landscape" details how IBM’s software reduced costs by pinpointing where the students who were likelier to stay at the university came from. "A set of equations weighing demographics, academic history, and other factors" were analyzed to identify which "have the highest probabilities of coming to Wichita State." Based on that, the university adopted a more targeted strategy for recruitment.

For example, after analytics revealed where the vast majority of the university’s students come from, the admission department focused on those high schools. The revelation that very few students come from outside the state prompted the university to cut 14 college fairs and reduce travel. They also took a more focused approach to their direct mail. In the past, they sent out 9,000 letters. After applying analytics, they only had to send out 5,000 to 6,000. The decreased number of letters actually translated into an increase in recruitment of 26 percent.

Preparing for Tactical Changes

In an email exchange, Wright explained the challenges of getting an institution to switch gears and embrace analytics. He said three aspects were involved:

• One was getting people to see the benefit of evidence-based decision making. Using data to make decisions is very different from using data to confirm a decision. In the beginning, the university had a hard time getting people to use data prior to the decision point. The data should be at the table as decisions are made.
• The second difficulty was getting folks to trust the analytics, especially when the data are so contrary to intuition or past practices. It took a long time for advisors to have faith in the data.
• And third was the quality in data necessary to use analytics.

In order to get a robust analytics system in place, they had to first clear out old data and "thousands of data entry errors." That was a daunting task, but the university agreed to it for the sake of setting up the robust analytics system that was necessary to achieve their goals.

Better Data = Better Employees

Applying big data analytics has also been proven to improve recruitment and retention of employees. Big data company Evolv is in the business of applying predictive analytics to hiring in particular. That’s because using big data to direct hiring decisions pays off, according to the company.

For example, Evolv’s insight changed Xerox’s hiring strategy for selecting call center workers. In a WSJ article, Xerox’s chief operating officer of commercial services admitted, "Some of the assumptions we had weren’t valid." That’s the real value of big data analytics; it reveals actual correlations based on objective information rather than gut feelings of hiring managers.

As it turned out, resumes and background checks turned out to not be the most reliable indicators of Xerox employees who would stay on until the company gets a return on its $5,000 investment in training. Evolv’s data showed that a record of arrest that dates back over five years does not indicate "future bad behavior" any more than a perfectly clean record. A previous record of job hopping also doesn’t necessarily mean the new hire won’t stay put. Evolv completed a study of 21,115 call center agents. The analysis of the data indicated "very little relationship between an agent’s work history and his or her tenure in the position."

What are the factors that do make a difference then? Personality, connections and location. Evolv’s software identified the ideal candidate as a creative person who is active on one to four social networks and is within a manageable commute of the workplace. Another key factor in retention was association. The ones who proved likeliest to stay at a company were those who knew three or more employees who already worked there.

Differences In School and Business

While big data analytics can be as effective in corporate recruitment as it is in university recruitment, it also shows where the parallels between the two break down. In a 2013 Forbes article, about what a company learned when it applied predictive analytics to selecting sales people, author Josh Bersin points out that school experience counts for far less than people think in terms of predicting job success. In fact, contrary to popular belief, a candidate’s GPA or choice of college did not correlate with success on the job.

That doesn’t mean that education is without value; finishing some form of education was one of the indicators of career success, but the key there was completion rather than the school or grades. Other key indicators included a grammatically correct resume, demonstrated success in a job, successful sales experience and the ability to work under unstructured conditions. After the company incorporated the data analytics into its qualifying steps and identified the factors that were accurate predictors, it improved sales performance to the tune of a $4 million gain in revenue.

Whatever the needs of the organization, predictive analytics can put them on the right track. As Wright said about his own experience, "By empowering folks with the resources they need to make good decisions, everyone wins."