Almost all institutions are using social media and email to communicate with students, potential students, alumni and other audiences. Those interactions are creating vast amounts of unstructured data—with rich potential for data mining, if you know how to do it.
“There are free to low-cost tiered tools available out there that can help people evaluate their presence on social media, specifically Twitter, as well as analyze the effectiveness of inbound and outbound e-mails,” said Jim Bouse, AACRAO President-Elect and Director of Enrollment Management Technology and Associate Registrar for Technology at the University of Oregon.
Over the last year, Bouse and Joshua Klenk, the university’s Enterprise Systems Integrator, have been analyzing trends involving their institution using IBM’s Watson analytics. The program uses natural language processing to reveal insights from large amounts of data.
“You can take a look at who is talking, what are they talking about, and measure certain keywords or hashtags that are important from an institutional perspective,” Bouse said.
For example, if your university was hiring a new president or engaged in another newsworthy event, the tool could sift through tweets about the subject to identify positive, negative and neutral responses. In addition to analyzing sentiment, the tool can also break down these reactions using demographic measures.
The tool can also be used with messaging campaigns, such as admissions office communication with prospective students. You can use the tool to analyze the tone of your email—is it more tentative or confident, assertive or passive?—and, using trackable URLs, determine what elicits the most desired click-through response (such as increased clicks to fill out forms, or decreased “unsubscribe” clicks). Then you can craft your message using the most effective language. And you can also compare the language analysis of your message to a similar message from a competitor.
“The tool is good to use alongside anecdotal evidence about what is and isn’t working,” Bouse said. Incorporating that information, you may then craft a message that will better reach a defined audience—such as a specific geographic region or a targeted G.P.A.—because you learn they are motivated by different messages or communication styles.
“We’re really looking at correlation versus causality,” Bouse said. “There are so many factors that you may not be able to nail down a single 'magic bullet' causal relationship. But you can get insights from correlation and think: ‘What kind of course corrections can we make? What behavioral changes are driven by those course corrections?’”
Bouse will discuss “Data Mining and Analysis Using Social Media and Interaction Data for Proactive Engagement” at this July’s AACRAO Transfer and Technology Conference in Anaheim, California. For more innovative CRM strategies, learn more and register for the conference now.