Database Marketing: Marketing Mavericks

• By Matt Villano
• 10/26/05

Whether it’s to dramatically improve enrollment numbers and quality levels, or boost advancement dollars, the latest analytic theories and database marketing strategies are making a huge difference on college and university campuses nationwide. And although Business Intelligence (BI) software is the newest find for many US schools, many of these insightful approaches go beyond basic BI tools, which are fundamentally designed to help technologists and administrators turn masses of data into sensible reports. Instead, these newer approaches provide more targeted methods of, say, attracting prospective students or communicating with alumni who donate most frequently. These newer tactics organize business intelligence in a way that benefits users across the institution.

Enrollment Modeling

Nobody said attracting the most qualified students was easy. The first hurdle is the cost of getting the “right” students interested— a process that can run into six figures before a single marketing pitch is made. According to a recent report by Marquette University (WI), for instance, the school recently shelled out nearly $200,000, or $3 per brochure, just to get its annual average of 65,000 brochures out the door. Compounding the problem of “finding” students is the rise of online applications and the preponderance of low application fees: More students are applying to schools than ever before, making prospect pools virtually unmanageable. To cope with these issues, many schools routinely resort to a funnel approach, buying tens of thousands of prospect names and mailing expensive direct mail pieces to attract interest. Yet, the challenge of sifting through endless data continues to daunt campus marketing professionals.

The newer enrollment modeling approach to database marketing identifies the key characteristics of enrolled students, allowing a school to determine the likelihood of prospects and inquirees to enroll. It’s actually predictive enrollment modeling that uses student characteristics to build a statistical model that assigns a score to each prospect or inquiry. With this approach, Lawrence Henze, managing director of Blackbaud Analytics, says schools can target their marketing efforts to those prospects who are most likely to respond, saving costs and ensuring admission departments meet their enrollment goals (see box). “The reliance on large numbers is reduced when there is a strategic enrollment system in place to guide prospects through inquiry, application, admission, and enrollment,” says Henze, whose company is a new division of the software firm Blackbaud. “Many schools may find they can cut their initial search purchases when they target prospects using enrollment modeling.”

How d'es such modeling work? Vendors such as Blackbaud, Datatel , SPSS, and SAS each have different products that boast proprietary algorithms, but in essence, all of the approaches revolve around a 1,000-point scoring system that grades prospects on the likelihood they will enroll. The scoring matrix awards points based upon characteristics such as geography, grade point average, and average household income (these characteristics vary by institution). Once the grades are assigned, prospects with high scores (generally in the range of 700 to 1,000) receive diligent communications, while prospects with lower scores become secondary priorities, and prospects with the lowest scores are dropped from the radar screen altogether.

According to Henze, higher education institutions using predictive enrollment modeling enjoy a number of advantages and cost savings. First, by targeting applicant pools, schools can identify the best enrollment prospects by geography, gender, ethnicity, field of study, academic standing, or other characteristics, targeting recruitment effortsto those students who are most likely to enroll. Second, he says, schools can save money on postage and printing by distributing communications more effectively, and save time and effort by making sure that admissions staffs are focusing only on those students who are likely to come to the school and matriculate.

Abilene Christian University (TX) has been using predictive enrollment modeling to assist in market segmentation to build and enroll better pools of students since 2000. According to a 2003 report published by the Journal of Marketing for Higher Education (vol. 13, issues 1 / 2) which compared Abilene’s first year of predictive enrollment with the previous two years, ACU received 25 percent more applications and increased enrollment by 10 percent, with between 22 and 40 percent more applications coming from the core segments the university wanted to grow. Bart Herridge, the school’s director of Enrollment Planning and Research, told the Journal he was delighted with the results. “We saw predictive modeling as an opportunity to create an inquiry pool shaped like we wanted,” he was quoted as saying. “[It] allowed us to leverage our costs while dramatically increasing the quantity and frequency of direct mail and telecounseling efforts to those students in our inquiry pool who were most likely to enroll.”

Theoretical Approaches

Marketing to alumni and prospective donors presents whole new sets of challenges to the campus marketer. In fact, entire technology segments have sprouted in past years to help colleges and universities tackle the chore of cleaning data so that they have the most current information for alumni. This segment, the “address reclamation” niche, is now well established, and vendors such as Market Data Retrieval, Cleanlist, Business Credit Information, and Melissa Data have been pivotal in helping schools manage their data. What is new, however, is the theoretical notion of applying standard Customer Relationship Management (CRM) standards to a group of alumni, treating these prospective donors more as long-time family members than as occasional friends.

Results of research in this area are surprising. At Columbia University (NY), for instance, Assistant Professor of Marketing Oded Netzer has developed a method to assess how an entire sequence of a school’s marketing actions, and other contacts with alumni influence alumni donating behavior. Using a sample of 2,000 alums from Stanford University (CA) over a period of 26 years, Netzer’s model determines the evolution in the relationship between the alumni and the university, the impact of development activities such as reunions, and the donating behavior of alumni in general. By establishing patterns of donation behavior from the donation database, Netzer has developed a new way to estimate donations, a schematic marketers can use accordingly.

“Previous methods were limited in their ability to bring together all customer- brand touchpoints, and therefore are less accurate in predicting long-term customer behavior,” says Netzer. “Applying this new method, [schools] may find that some of [their] marketing actions have a more enduring impact on [their] customers, which in turn implies a different marketing strategy.”

As Netzer explains, his method takes CRM to a higher level. Using the wealth of data from these programs, the Netzer approach lets schools a) segment their alumni according to what kind of relationship they have had in the past, b) determine how the relationship changed over time and where is it heading, and c) calculate which marketing actions are most likely to move alumni to a higher level of relationship and donation. The approach also provides a marketing strategy decision-support system that predicts the effects of different actions on different types of customers, over time.

Embracing the notion of taking a broader approach to database marketing is On Amir, a professor of Marketing in the brand-new Rady School of Management at the University of California-San Diego. Like Netzer, Amir emphasizes the need to optimize the types of data a school is keeping on its constituents. In a candid interview, he says that for the greatest return on their investment (ROI), schools engaged in database marketing always should ask themselves the following questions: How old are constituents? What are they interested in? Do they prefer personalized messaging? How many different types of groups exist? Amir notes that with more advanced databases, schools could even attempt to assess emotional and personal attributes such as what time of year an alumnus is most likely to donate, or what kind of attachment a prospective student feels for the institution, from the beginning of the campus/ student relationship.

“The trick in marketing is that you want to create the most value you could have, with long-term ROI,” he says. “That value increases as you get more specific about your data and your marketing groups.”

What’s Next

Looking forward, database marketing experts see challenges ahead for colleges and universities that plan to leverage data to get more of a return on their marketing dollars.

First, schools using predictive enrollment modeling or more adventurous approaches like Netzer’s are few and far between. Second, while a majority of colleges and universities rely on BI software to report historical data, the online analytical processing (OLAP) component of this software is frequently illequipped to guide prospectivedecisions.

But another big issue with BI is where to put the software programs to make them most valuable. David Wells is director of Education at The Data Warehousing Institute, which provides education, training, certification, and market research for executives and business intelligence professionals worldwide (TDWI is a sister organization of this publication). Wells says academic technologists can’t seem to figure out whether business intelligence programs should report to IT, university administrators, or the academic side. The most successful BI engages the business, and is a “business first, technology second” endeavor, he explains, adding that the optimal structure removes the organizational separation of business and IT for those engaged in business intelligence work, creating an organization where business knowledge and technical knowledge are contained within the same business unit and frequently within the same individuals.

This approach d'esn’t have to involve major restructuring; on the contrary. Wells suggests that schools create the BI organization as a federation of skills and knowledge, with everyone working toward the same goals, the same purpose, and with shared accountability for effectiveness and value.

“In typical universities, information technology skills are found both centrally and dispersed throughout academic and administrative departments,” he says. “It’s best to comprise a federation of people, knowledge, and skills that encompass academics, research, university administration, and IT.”

Another challenge for higher ed institutions pursuing database marketing is applying the concepts of Enterprise Data Warehousing from the start, recognizing that “one size fits all” analytics for the entire institution might not be the most efficient way to approach a database. Here, Wells notes that most large universities are not single enterprises, but conglomerates of many. Technologists at the University of Washington understand this reality perfectly. In addition to the undergraduate and postgraduate programs, the school runs two hospitals, and is the second largest recipient in the nation of Department of Defense research grant funds. With this in mind, the school actually operates as three distinct enterprises— education, research, and health care. Each “enterprise” has its own data warehousing effort, which means that each also takes responsibility for its own database marketing.

“When it comes to database marketing, particularly in higher education, half the battle is making sure that you’ve got enough control of your resources to make it worthwhile,” Wells observes. “What you do from there is up to you.”