What is the benefit of utilizing a predictive model in higher education?

  • The benefit of utilizing a predictive model depends on the type of model created. If a recruitment model is built, whether it is an Inquiry or Application model, the intent is to understand a prospective student’s likelihood of enrollment. Understanding this business intelligence allows an enrollment manager to make better marketing and resource related decisions. For example, understanding that a recruitment model ranks 94% of all eventual enrolls within the top 50% of all inquiries allows an enrollment manager to prioritize and/or segment communication strategies which in turn could reduce mailing cost. It could also help with managing the recruiting staff by more effectively scheduling travel or more appropriately delegating call times. In the case of an application model, it can assist in the prioritization of admission interview calls. If a retention model is built, the intent of the model is to understand an enrolled student’s likelihood to not persist to a future term. Understanding this information allows an enrollment manager to identify ‘at risk’ students as early as possible to develop executive interventions and to better understand factors leading to attrition which in turn will lead to improved retention rates.

  • Who can benefit from predictive models?

  • Four year Private, Four year Public, Community and Technical Institutions can all benefit from predictive modeling by applying them to their traditional undergraduate, graduate, adult, online and other program types. As long as an Institution maintains sufficient and accurate information about prospective students throughout the enrollment funnel, they can benefit in some way from a predictive model.

  • What type of model should a school look to implement?

  • The type of model really depends on the current goals of the institution. If the institution has a large and potentially unmanageable volume of inquiries in addition to a poor inquiry to application ratio they would probably benefit most from an inquiry model. However, if the institution is satisfied with their application rate but the application volume is large and unmanageable they may benefit more by implementing an application model. Alternatively, if the institution is concentrating more on retaining enrolled students rather than recruiting new ones then a retention model is the logical choice.

  • How long does it take to develop a working model?

  • The creation of a model is greatly dependant on the state of the data provided. Assuming minimal cleansing and no data structure changes are required; a model could be built and implemented within 4–8 weeks.

  • How much historical data is really necessary to build a good model? What if a school doesn’t have all the data?

  • Typically we look to utilize the most recent enrollment term for building a model. Depending on the type of model and/or the volume an institution experiences we may look to expand to 2 or even 3 years worth of data. However, we always prefer to limit the model build to one year as there are many factors that influence enrollment and retention decisions that can change from year to year. If an institution does not currently have all the data required to build a model, we suggest working with someone knowledgeable about predictive modeling in Higher Education to help develop a plan for collecting and maintaining the data necessary for building a model upon completion of your next enrollment term.

  • Can one school use another school’s Predictive Model?

  • No, the models that we build are customized to each institution. We believe strongly that the best supporting data for any enrollment model is the institutions own enrollment data and putting in the time to use such unique inquiry, application, or financial data is a major reason why these models perform so well.

  • How is a retention-focused predictive model different from a recruitment-focused model?

  • A retention-focused model differs from a recruitment-focused model in that the intent of a retention model is to predict, out of a group of enrolled students, who is most likely to not persist to a future term. The intent of a recruitment-focused model is to determine, out of a group of inquired or applied prospective students, who is most likely to enroll in a future term. Each type of model is built and applied to a different stage of the enrollment process and utilizes data available at each specific stage.

  • What should be included in a retention model that is not included in a recruitment model?

  • The most important addition to a retention model, that is typically not available for building recruitment models, is information submitted on the FAFSA and financial aid related information.

  • If a school is interested in Predictive Modeling, what can they start doing in anticipation/preparation?

  • They can begin placing a priority on capturing and maintaining high quality data for analytic purposes at each stage of the enrollment process. They can also place an emphasis on collecting consistent data across all channels that they may use for generating inquiries.

  • What are some characteristics of a successful model?

  • A successful model should provide a meaningful discrimination between those that enroll (or persist) and those that do not enroll (or do not persist).

  • What have you seen successful schools accomplish with their predictive models?

  • In general, we have most often seen schools create more efficient and cost effective recruiting practices. In some cases we have seen the institution reinvest cost savings to enhance communication to higher ranked prospective students and in other cases we have seen them use the model to cut costs and better organize staff responsibilities. We have also seen an institution use their model to identify a manageable group of inquiries to include as part of an in-house telemarketing campaign. In the case of retention modeling most often schools use the model as an early indication of ‘at risk’ students. Using this information they target retention efforts by increasing retention support for high risk students through assigned success coaches and carefully monitoring the behavior of average risk students.