Even in these digital times, undergraduate admissions remains a paper-laden discipline. Viewbooks, search pieces, postcards, catalogs, applications, and more need to be printed, enveloped, and mailed, a process not only costly but also inefficient. Most inquirers to any one school, after all, end up attending elsewhere.
In the late 1990s, faced with state budget cuts, the admissions office at State University of New York's Alfred State College began searching for ways to reduce printing and postage costs without affecting enrollment. It had been sending more than 100,000 pieces of mail to 17,000 potential students, of whom fewer than 10 percent—around 1,500—enrolled.
Alfred State's solution was predictive modeling. Administrators purchased a target statistical analysis package (from the now-defunct Target Marketing) that reviewed the previous three years' inquiry, application, and enrollment data and created a profile of students most likely to enroll as freshmen. The program determined that between 80 and 90 percent of enrolled students shared a profile common to only 40 to 50 percent of applicants.
Admissions staff began uploading new inquiries into the software, which analyzed various components of applications and assigned them ratings reflecting the inquirers' likeliness to enroll. Those rated more likely progressed through the college's typical six-mailing cycle, while those deemed less likely were mailed to only three times.
The predictive modeling package enabled the institution to target admissions mailings to its best prospects.
Alfred State was able to reduce the number of mailings by 25 percent, saving more than $20,000 a year in postage alone and also realizing savings in printing and stationery. The institution realized efficiencies of time as well: The ratings were used to prioritize the outreach activities of faculty members and student ambassadors.
"There are students who may express an interest and no matter what you do are probably not going to enroll," says Deborah Goodrich, associate vice president for enrollment management. "... You can't be all things to all people, and you can't spread all of your recruitment initiatives equally among all students."
Goodrich describes the initial decision to reduce mailings to some applicants as "a risk," but notes the college took steps to mitigate that risk. After implementing the package, the admissions team operated according to business as usual for the first year and ran the system as a sample to gauge its predictive ability. Only after determining its accuracy did Alfred State begin altering its mailing process.
The college updated the package in 2001 to reflect the addition of several new academic programs, since an applicant's choice of major is a significant factor in his or her decision to enroll. It was updated again last year, though with Target Marketing no longer in business, it needed to go through a new vendor, PlattForm Higher Education. The system now includes three top-level codes—for best, good, and doubtful prospects—with dozens of gradations within each level.
"It surprises me that this is something that is not used more widely by my colleagues," Goodrich says. "The success we have found is just phenomenal. ... When your resources are limited, you can really plan how to direct them.
(More information on PlattForm Higher Education can be found here.)