Main Nav

EDUCAUSE Analytics Sprint, Day 3 Recap: Analytics for Enterprise Efficiency and Effectiveness


Thursday, the third and final day of the EDUCAUSE Analytics Sprint. Today we heard about enterprise-level activities to support analytics at the University of Washington (UW) and at Arizona State University (ASU). By now, everyone knows that analytics programs depend on good data, and we all know that at most institutions, ensuring appropriate access to reliable data is not trivial. As was noted today, colleges and universities are “loose confederations of guilds” in which people tend to fiercely guard their data.

Enter data governance. A poll of Sprint participants showed that just over half have not even begun a data-governance program. If you think that effective data governance can be dizzyingly complicated, look no further than UW. With its 17 schools and colleges, nearly 50,000 students, and 50 institutes and research centers around the world, UW is enormously complex, though it is hardly unique among large institutions. If you think that wrangling all that data is an impossible task, though, look no further than UW, which has created a governance structure that encompasses standards for data quality and access, user roles and responsibilities, processes and jurisdiction, lots of acronyms, and many other aspects of data stewardship. Data governance isn’t likely to be easy, fast, or cheap for any college or university, but with an institutional commitment, it can certainly be done and will surely have benefits beyond an analytics effort.

With more than 72,000 students, ASU has even more people producing data that can be captured, stored, and analyzed. Presenters from ASU noted that data has become a strategic asset of the university, and the institution has invested heavily in drawing value from those data assets. ASU has built sophisticated systems that provide insights at differing levels, from technical tools (that the rocket scientists like) to dashboards for, um, the rest of us. ASU focuses considerable effort on the ways in which data (and patterns) can be easily consumed, not just by faculty and administrators but also by students themselves. This gets to one of the themes of the Analytics Sprint: It’s easy to imagine an analytics program that falls so deeply into maintenance and policies and applications that we become slaves to the data, but analytics has to be designed to make data work for us.

Factors including institutional size, the degree of decentralization, and the age of computing systems affect the complexity of getting data into shape for an effective analytics program, and both UW and ASU have devoted considerable staff and other resources to their efforts. All of this might sound daunting to many institutions, and perhaps rightly so. But what if your institution is much smaller, less intricate in its structure, with a history of better-than-average data stewardship? Maybe your institution could far more easily implement an effective analytics program. Maybe.

Regardless, the Analytics Sprint made clear that analytics involves work on a long list of fronts, including policies, governance, technical apparatus and tools, training, data access, privacy and security, and data quality—not to mention the evaluation and ongoing maintenance of these. For those willing and able to make the commitment, however, the results can be striking.