Driving the Data
DATA-EVERYBODY HAS IT, but they aren't all using it the same way. By applying business intelligence (BI) to reporting, some administrators are taking data analysis to the next level. "BI provides a qualitatively enhanced relationship between the decision maker and the overwhelming amount of information created on a daily basis," explains Keith Myers, VP of the Technical Services Division of SunGard Higher Education. Successful BI helps decision makers reach a state of data exploration, integrate data across time, and uncover hidden relationships, he says.
"We've embarked on a new way to look at what we do and how we do it," says Marsha Drennon, president of State Fair Community College (Mo.). Administrators are in the process of implementing an Operational Data Store and Enterprise Data Warehouse from SunGard as part of their overall BI strategy.
Drennon's institution had to change to a more data-driven model after being approved for the Academic Quality Improvement Program from the Higher Learning Commission. "Our efforts at this new accreditation process focused on intense professional development, student centeredness, student success, and service to our community," she explains. Because the program requires institutions to work on four or five action items at all times, data-driven decision making is essential.
Once the new data systems are in place, administrators will be in a better position to analyze the information and decide whether any processes need to be changed, adds Michael Ash, vice president for student services at State Fair. The executive team now has eight key performance indicators they use as benchmarks and for comparison to other institutions.
"Community colleges are increasingly using data to manage persistence and retention outcomes and reporting on their success to the community," says Graham Tracey, director of higher education services at ARS Analytics, a consulting service. He points to accountability initiatives such as Achieving the Dream and a need to establish standards of success separate from those used by four-year institutions as driving factors for adoption.
"[Being data driven] makes us think about what we do and why we do it," says Drennon. Although implementing BI can be complex-the migration has taken about two years-she believes it is worthwhile. "Our trend lines are changing, and we can relate that back to what we're doing. When we run the reports, seeing where we've been and where we've gone is very exciting."
Although BI can be a powerful tool, it is not widely used in higher education. According to a recent study by Datamonitor, an online data, analytic, and forecasting service, across all levels of higher education only 29 percent of institutions surveyed currently use BI for customer intelligence, which would cover recruiting and retention. Fifty-one percent of those surveyed intend to have it in the next six to 24 months.
With the enormous pressure community colleges are under to increase enrollment while keeping costs down, they are positioned to quickly benefit from BI. "BI applications are providing the evidence for market needs and the evidence for the alignment or misalignment of current institutional investments and programs with market needs," says Bill Graves, senior vice president of Academic Strategy with SunGard. "Community colleges, out of necessity, are urgently charting the path towards public higher education's future with the aid of BI tools and the culture of evidence and decisive action that they enable."
"We're looking to be able to not just have the data, but ascertain causal relationships," explains Jeffrey Rafn, president of Northeast Wisconsin Technical College. "If 'X' exists, something will occur."
As an example, Rafn explains that the monitored, self-paced, open-entry business labs at NWTC had a 54 percent retention rate because some students were signing up in order to maintain 12 credits, allowing them to keep the college's health insurance. BI lets administrators track the trend and predict which students might become repeat offenders. Now students who don't complete two labs have to see an advisor before they can sign up again.
Administrators at Des Moines Area Community Colleges (Iowa) are also planning to use BI to predict student success. Joseph DeHart, executive director of institutional effectiveness, has identified 27 variables-including how long individual students waited to register for classes after they were accepted, whether they are first generation, and testing results-which might not be informative individually but are powerful when examined together. Using software from SAS, administrators at the school will be able to determine which variables have a bearing on student success. "You will get a profile to target successful students and also be more aggressive with intervention," DeHart says. He is still working on the models and creating ones for suburban, urban, and rural students in order to prevent profiling.
"Reporting is entry level," DeHart says. When he sat down with DMACC recruiters, they all requested different information. "They didn't need reports. They needed data sets," he explains. With the SAS software, DeHart can ensure the data is correct however people manipulate it.
Rafn had a similar experience when looking into retention at NWTC. Reports showed how many students were getting a C or below, which also showed administrators how many students were passing but did not address retention.
"People are used to seeing up-to-the-minute statistics. Those are descriptive statistics. They show what is, not why," says Tracey. "In a good BI regime, that data gets transformed to create views of the data that might not be considered at the moment."
Tracey says it's important for BI users to ask themselves what they are going to do with the information once they have it. He also suggests starting with a small project for which they already have reporting, in order to see the benefits of BI sooner.
By examining data collected over three years through the Student Satisfaction Inventory performed by Noel-Levitz, State Fair administrators found that the largest gap between expectation and student satisfaction was in advisement. While advising is required of first-year and transfer students, Drennon says that institutional leaders are expanding mandatory advising to the entire population. A pilot program with 500 students has already been completed.
"Our data is two semesters old, but students are seeing it positively, and our retention rates are up 2 percent," Drennon says.
The most important part of BI is getting people to ask the right questions, notes Rafn, whose institution receives on-site technical support from SunGard. NWTC employees are receiving training that will help build a culture of evidence on campus. All constituents also have the ability to run ad hoc reports to find the information they need. After all, Rafn points out, if people have to go through the IT department, "they either don't ask or they create something outside the system."
DeHart echoes the sentiment. "We're trying to get away from the burden of one person running all the reports." End users at DMACC can create and save data sets on their own. And DeHart's starting to see a culture shift. "In higher ed, if you ask people what they want they won't know, but if you give them something to respond to, it gives them ideas. They'll see a report and ask for budget data too."
But being data driven isn't easy or inexpensive. "It complicates our lives to a degree," Drennon says, adding that making an investment up front is critical to staying competitive and retaining State Fair's niche, since students have many educational options.
Rafn says that at his institution, BI software cost $200,000 but that it was well worth it: "If you improve retention by 5 percent and you have 6,000 students, ... you've more than made up your cost." He says when it comes to being hit or miss, BI increases the chance of a hit.
"Comparatively, our BI isn't any more expensive than other systems we run," says DeHart. He says institutions considering BI should have needs and objectives firmly in mind before talking to vendors. Once a system is selected, he also advises allowing time for training and for users to learn the system. But most importantly, he says, "there is no substitute for knowing your data. You have to know the data in your SIS inside and out. BI just lets you leverage that knowledge."