There is a lot of buzz about analytics, particularly predictive analytics, in industries like higher education. In higher education, tight budgets mean every dollar spent on analytics better pay dividends. Universities and other institutions all recognize that to compete in the digital economy, they need to be more efficient and innovative in their use of analytics. For example, they’d like analytics to predict the likelihood of common events such as students most likely to enroll and those at high risk of dropping out. Other analytics uses include uncovering strategies for improving the student experience or evaluating the giving potential of certain donors. All of these use cases and more allow institutions to spend wisely and act strategically. In fact, institutions that hesitate to use the power of analytics to increase efficiencies, lower costs and gain competitive advantage will soon be left behind.
While institutions are eager to reap the rewards of these new methods for predicting behavior, they can easily be disappointed in their results if these programs aren’t built on a solid foundation. Just as an expensive house should never be built on sand, it is also fiscally irresponsible to build an analytics program without a solid data management foundation. The problem is, many institutions do not know what a solid foundation for analytics looks like. To help, the following are five critical attributes of any successful analytics foundation:
1. Develop an Overall Data Management Strategy: Analytics should be one part of an overall data strategy. If your institution has a Chief Data Officer (CDO), then they should take responsibility for planning and executing an enterprise data strategy, and fostering a data-driven culture. If not, someone else will need to fill this role. In any case, someone needs to take ownership of developing an overall plan, and not just from a technology or architecture perspective. While having a cohesive architecture strategy is critical, it is just as important to formulate a business strategy that considers how business needs are expected to evolve over the next 3, 5 and even 10 years. This broad strategy can be developed internally, or with help from consultants with expertise in assessing their customers’ business and data challenges, needs and goals, and developing a roadmap to achieve desired outcomes and objectives.
2. Establish a Data Governance Framework: Data governance programs help data consumers answer critical questions about the data available to them. This includes users creating business intelligence (BI) reports and analytical models. Some of the questions a data governance program helps address are:
Creating both a central repository for these answers and establishing consistent information ensures easy access and common understanding among users. It also removes barriers of knowing who to ask and the risk of having people with unique institutional knowledge of systems or processes leave the organization with it.
3. Ensure Data Quality: In order to build on a solid framework, the data used to create predictive models or any analytics needs to be accurate, consistent and reliable. This can be accomplished by utilizing balancing and reconciliation techniques as the data moves between systems and processes. This includes not only ensuring that each transaction is successfully moved between systems, but that any critical pieces of data are transformed correctly. In addition, a number of validations should be performed on the data to confirm that the data is complete, consistent, conforms to intended formats and is not duplicated. Data Quality is a key piece of building an analytics foundation that many people either overlook or underplay the impact of doing this well.
4. Empower Users to Transform Data: There are currently two ends of the spectrum when it comes to data transformation. On one end, IT does data transformation at a large scale using traditional ETL tools. At the other, business users perform one-off data transformations using spreadsheets. While both ETL tools and spreadsheets have a place in an organization, there are a number of functions where they’re a less-than-optimal tool for the task. Examples include merging data sets, renaming fields, transforming data with computations and other ways of preparing the data for BI and other analytics. Not only are these tasks best performed using a tool specifically designed for that purpose, this allows them to be easily automated for repeated, ongoing use and easily shared with other users.
5. Automate Action: Once predictive models have been developed and run, what do you do, or what should you do with the results? As part of your overall analytics strategy you should have determined what action your organization will take based on the outcome of each predictive model. If the sentiment of current students regarding satisfaction with their current university experience is trending down, what are you going to do about that? Are you going to send them an email or have someone call them? Perhaps you want to identify any students for whom the sentiment analysis shows that they are 10% less happy today than they were last month. You should automate the execution of the model and subsequent notifications to those who can take action. In this example, the director in charge of outreach to these students could receive a prioritized list of those who exceeded the defined threshold, so that outreach can begin immediately.
As you can see, while analytics and predictive modeling are critical in higher education, there are a number of foundational pieces that must be in place in order for the predictive models to be effective. We all want to ensure that we’re spending our limited resources as efficiently as possible. The only way to get the most out of your analytics budget is to ensure your analytics projects are built on a solid foundation of a business strategy layered with programs for data governance and data quality. It must include best practices in self-service data preparation, and most important, automate the actions you want to see as your result. This will maximize not just budget dollars, but the organizational impact of analytics efforts.
Are you looking for an enterprise data management strategy to bolster analytics in higher education? Check out the webinar below.
For a deeper dive into this topic, visit our resource center. Here you will find a broad selection of content that represents the compiled wisdom, experience, and advice of our seasoned data experts and thought leaders.