• Special Feature / Interview
  • 03.16.21

Strategies for Using Predictive Analytics

  • by Georgia Mariani, Principal Industry Marketing Manager, SAS, Higher Education
Karl Konsdorf

I interviewed Karl Konsdorf, Director of Research, Analytics and Reporting at Sinclair Community College, to learn about their strategies for using predictive analytics.

Can you share some examples of how you are using Predictive Analytics? How will you put the analytical results into practice?

We are analyzing student behaviors and doing predictive modeling to better route students to our two-tiered advising system – counseling or advising. Counseling is much more intense and expensive, so we want to make sure that the students that receive the counseling are the ones who will benefit the most from it versus just academic advising. This will help counselors as they are focusing on those students that will need this level of assistance. In addition, this will help students because we make sure they’re getting the right support without overwhelming them with intense counseling.

Another project involved working with advising to determine the most critical students that need assistance. Advisers have about 200 active students and a caseload of 600 total students. Using predictive analytics, we were able to determine the students that will benefit the most and provide an advisor with a top 10 list of those students. Using these lists, advisors keep in contact with those students to get them registered for next term.

Why is this important to your organization?

In the past, decisions about routing to advising or counseling were based on using reactive data, like 1st term grade point average. We really wanted to be more proactive by leveraging intelligence on the front end instead of intervening after the student could have benefited from the services. By using analytics and being more targeted, counselors will be able to work with students that need a deeper level of assistance, and advisors can then concentrate on all the others. As such, we rely on technology to automate the routing decisions.

What advice do you have for other colleges that are considering using analytics?

Have a plan for the distribution and the consumption of the results. Our best practice is to integrate the analytical results into the end-user application, so that it’s viewed as part of the business processes. For example, in the advising/counseling model, our plan is to incorporate the results directly within the case management application as an indicator that routes the students appropriately.

Can you share advice on building an analytics community or culture?

I see four practices for doing this.

  1. Organize for success. I am a strong believer in the Business Intelligence Competency Center or Center of Excellence Model where you incorporate the institutional research, data management, business analyst into one entity. This is critical to success as it eliminates obstacles in developing, deploying, and consuming analytics.
  2. Deliberate data management strategy. Define, plan and layout your data architecture. Our mission and vision are to be the single source for decision support for the college, and our mission is to build an integrated information architecture from student and business activities to support decision making processes. We know how we're going to handle data quality, how we're going to model data for reporting, modeling for analytics, and how we're going to manage our predictive models.
  3. Ensure content is simple, standard, and significant. The three S’s, I like to call them. First, content should be simple to use and easy to understand. Next, variables should have standard data definitions across reports. Finally, reports are significant - meaning that they are pertinent to the audience that's using the information.
  4. Engage the community. For the last four years, we have hosted our annual Data and Completion Summit that is well attended by faculty, administrators, and staff. They attend hands-on workshops on how to use some of the new dashboards. They learn about different programs for completion and how they're using data to measure success in their programs. They also learn how to engage with research staff to use either institutional review board, how to ask simple research questions, and how to work with survey data.

Without the use of analytics, where would your institution be?

Left behind. That's where we would be.