• Better understand the relationship between key student variables and retention;
• Enhance our ability to predict which students have a high probability of leaving;
• Target interventions to at-risk students and evaluate interventions on their effectiveness in improving student success; and
• Enhance our ability to predict full-year equivalent and headcount enrollment.

The Enrollment Analytics program was deployed fully with the fall 2013 new-entering freshman cohort. These students have been tracked through a student success dashboard on the four student success outcomes that our research indicates are predictive of future student success (degree completion). We have seen improvement across three of the student success outcomes:

• Retention: 3% increase in first-year retention
• Credit completion rate: 1.6% increase in credit completion rate
• GPA: 0.13 point increase in first-year GPA

Our experience implementing enrollment analytics over the last few years has revealed the following challenges and considerations:

• Analytics work is iterative, with each analysis bringing new findings that sometimes require reinvestigation of earlier analysis. Models are built based on best available data and will need to be reanalyzed over time as new information becomes available, as policies and processes change, and as predictive models are tested against actual performance.
• Analytics requires extensive data cleaning. Data obtained through query-based processes or download are usually dirty. There will be missing, incorrectly entered, and misplaced data. Extensive patience and careful attention to detail is required to correct for inaccuracies and to provide correct interpretation.
• Analytics will bring to light existing organizational processes that are ineffective or debunk organizational myths that reinforce the status quo. As data is brought to bear on these issues, organizational leadership must be prepared to take action. 

Many tools and technologies are available to support campuses in their implementation of data analytics. SCSU uses MicroSoft BI tools, and the institution approaches data analytics through a five step process that can be replicated at any institution, regardless of the tool used:

• understand the processes you are attempting to improve and use data to test the assumptions embedded in your practice;
• assemble a cross-functional team that understands the process and the data;
• build the best data set possible but don’t wait for perfection;
• use data to tell the story of what is occurring; and
• be explicit about the organizational action that needs to be taken. 

Wishon, G. and Rome, J. (2012). Enabling the Data-driven University. Available from: www.educause.edu/ero/article/enabling-data-driven-university

Hrabowski, F.A., Suess, J. and Fritz, J. (2011). Assessment and Analytics in Institutional Transformation. EDUCAUSE Review, Sept/Oct 2011, p. 15-28.

Bichsel, J. (2012).  Analytics in Higher Education: Benefits, Barriers, Progress, and Recommendations. Available from http://www.educause.edu/ecar