Recruitment & Retention

Small BI Investments with Big Payoffs

• By Linda L. Briggs
• 05/13/10

Tiny Notre Dame College in Cleveland, OH tripled its enrollment--to 900 full-time students--over three years with help from a business intelligence solution that supported its information needs--and for less than $60,000 in software and consulting costs.

Smaller schools may see business intelligence as a big investment, but that needn't be so. In this interview, Christina Rouse, who has spent almost 20 years working in higher education as a researcher, faculty member, and, now, as a BI consultant, explains why spending top dollars for BI software isn't always necessary. "I think most small schools can do a really nice job of business intelligence for less than $100,000," according to Rouse, chief architect at Cleveland-based Incisive Analytics.

"For a smaller school," she said, "think of BI as costing in the tens of thousands [of dollars] instead of the millions."

Here, she offers several examples of institutions that have completed noteworthy BI projects for reasonable prices tags, especially focusing on recruiting and retention efforts--and have seen quick returns on their investments.

Linda Briggs: Do smaller schools tend to struggle more initially to get a good BI program in place?

Christina Rouse: They do struggle, often because there's an illusion that BI solutions are very expensive and time-consuming, and you have to have millions to get good business intelligence.

I approach the problem of BI in small schools by using what they already have in place as much as possible. Schools are usually simply looking for a different way to look at their data; they're usually not looking to collect new data, but to analyze existing data in a different way. So we often look at systems already in place to see what we can squeeze out of them. It boils down to buying some services, as opposed to buying hardware and software.

Briggs: What does a smaller institution usually have in place for its BI tools?

Rouse: They often have tools that they wouldn't think of as BI, but that can be used for business intelligence. The typical scenario at a small school is to have two types of tools in place. And by small, I mean perhaps fewer than 2,000 or 3,000 students, and as small as 500 or so students.

They often have a very traditional desktop environment, predominantly Microsoft-centric, perhaps with some [Apple] computers as well. But, it's very much Microsoft Office--Excel, Access, Word, and so forth. People are very accustomed to those tools, and they want to be able to continue to work in them. The school often has a campus-wide license at an educational price, which makes that part of the solution very affordable.

On the other end is the system that captures all of the data to begin with; that's usually an ERP system, something that may be on its last breath of support. It may be something purchased by another vendor and now being supported in a different way. It's usually not one of the three big ERP players for higher education--PeopleSoft, Datatel, and SunGard [Higher Education] Banner.

So, we want to be able to squeeze everything we can out of those two environments. Often, the schools have a Microsoft Campus Agreement, so they can go ahead and expand the license to include the business intelligence tools. Often, they don't realize that.

The other alternative is open source [data warehousing and] BI tools, such as those from Greenplum or Pentaho. Some of those are free or cost very little. Again, the institution doesn't need many licenses, just enough to get started.

So small schools generally have two choices: They can go with open source, or they can go with whatever their desktop is now, usually Microsoft.

Briggs: So these schools have often already collected the data, and they have some tools to examine it--they just haven't put the two together.

Rouse: They have the tools; they have the data; they just don't have the blueprint for the design. That's the piece that's missing. These smaller schools typically don't have staff with the needed skills. They have people that conduct reporting, and they have people that support IT, but they don't have someone who knows about BI design and analytics.... That's where a consulting firm comes in.

Briggs: What advice would you give to someone at a smaller school who knows the value of BI--perhaps has come from another school that used it--but is having trouble convincing the dean or others?

Rouse: It's always good to understand the value proposition of BI. If I were that person, I would look for one significant problem or challenge that institution faces and try to align my business intelligence solution to that problem.

For example, a lot of small schools with selective admissions have a huge number of applicants, then a smaller acceptance rate, and then a yield--those students that actually show up at the front door on the first day--that's even smaller. There's a lot of money tied up in managing that traditional admissions funnel, and you don't really know what your freshman class is going to consist of until that first day of fall term.

That's a very risky business. If we added business intelligence to that process, we could look at historical data and begin to take some of the risk away by predicting what's going to happen before it happens.

For example, if we have 12,000 applicants, accept 75 percent of them, but have only 12 percent show up, that's a lot of time and money spent on almost 9,000 students who aren't going to matriculate. If we know that ahead of time, can we spend less time and money on the 9,000 that we know aren't going to show up, and more time on those students that are. It's changing the slope of your admissions funnel to take some risk out.

That's where the money comes in. Admissions is directly tied to tuition... [A good BI solution] addresses both sides of the equation, the revenue and also the expense sides, in terms of tuition discount.

If you can tie business intelligence to [a clear example like that], it's a much, much easier sell.

Briggs: That's a great example. Is there another one you can share as well?

Rouse: Let's take a small school that functions more like a community college, in which people attend for certain pieces of their education, then leave. So the school has a high stop-out, high dropout rate. In that case, we want to look at retention. There's an assumption in some small schools that a student will be retained from first to second term. But if your school is a stop-out, dropout model that a student is attending to pick up a certain set of skills, the retention factor is huge, and the graduation factor is an even bigger challenge.

So if we could monitor first-term and first-year retention over a student profile and a mix of courses, we could begin to counsel certain students, telling them that "This combination of courses in your first term has a much higher chance of success in outcomes than a different mix of courses." We can also distinguish the person who didn't return after the first term because they truly completed what they wanted to achieve, versus the student who did not complete but simply didn't return the next term. That mix of courses is critical, particularly if you're offering developmental education courses.

One example I recall is a developmental reading course in concert with a psychology 101 course. That's not a very good mix because the Psych 101 course requires lots of reading. If we were to mix the developmental reading course with some courses that were more hands-on and activity-based, the chance of student success is much better. We learned that by looking at the mix of courses and student GPAs their first term.

Briggs: What kind of tool did you use to dig out that level of information from the data?

Rouse: The data needs to be stored at the student and the course section level. The tool to get at it could be just a Microsoft Excel pivot table.

Or, there are half a dozen tools on the market that allow you to look at what's called a market basket analysis. What courses did I have in my basket? Say I want to look at a cross-grid of everyone who took Psych 101, and everyone who took developmental reading at the same time. I have to be able to look at course mix by course mix. It's a traditional grid often presented in a pivot table [in a spreadsheet]--not an expensive tool. The challenge is capturing the data at the grain you need, and then organizing it in a way that allows you to see what the information is really portraying.

Briggs: As a BI consulting firm, once you come in and you set things up for a school, what happens? Does the institution still need consultants to help? Who becomes the technical expert once you're gone?

Rouse: That's a good question. Most organizations automatically look to IT to be the center and the hub of the business intelligence solution, and, while IT plays a part, their role pretty much stops at, "Are we legal [in terms of licensing], up and running, and secure?" We look at IT as the backbone of the BI solution because they're going to make certain the server is up and running, and it's backed up and recoverable.

Then we turn to other areas of the campus; I like to think of three different types of people in three departments: institutional research, where most of the compliance and the state reporting happens; the financial analyst, usually in the finance-HR area; and someone in financial aid because that's the person doing a lot of number-crunching around aid packages.

In that group of three types, you tend to have analytical people with technical skills, although they're not sitting in IT. But you need to look at institutional research, financial aid, and the financial services area for a couple of technical people that are good at writing reports. They understand basic SQL statements and how to turn data into information.

If you take one person from each of those departments and add a little bit of IT, you have a BI team. It's the little engine that could; it works very nicely. None of these people need to be full-time, by the way. You could have five or six people, and end up with perhaps two or three [full-time equivalent positions] to get your BI solution up and running.

Briggs: So schools often have the people already, along with most of the skills they need. They just need to give them a little time and training, and put it all together.

Rouse: Yes. They also usually need to buy a blueprint or design. It's very unlikely that you have someone on campus that understands the design needed to implement a BI solution. You would go to a consulting firm for that, but you should be able to buy that design and leadership to get you started. Usually in under six months, you can have your BI solution up and running, the need for consulting disappears, and you can support yourself.

Briggs: How soon might an institution see a return on that initial BI investment?

Rouse: It could be as soon as a couple of months. For example, if we tackle the admissions funnel problem, there's a particular application, the FAFSA, the Free Application for Federal Student Aid. Many, many students fill this out, after which the institution gets a copy of the FAFSA that says, "I'm requesting aid from your school." There's lots of information in the FAFSA that can be aligned to your admissions data.

One of our clients looked at the FAFSA and discovered that 1,000 students every fall indicated on their FAFSA that they wanted to attend the institution, but never filled out an admission application. That's a long list of potential students. If that organization netted just six additional students, it could completely pay for its BI solution. If an institution could get perhaps 10 new students, in many cases, it would pay for the solution. It's all about the value proposition of the data.

Another institution I worked with regained their investment in BI in seven months by changing their payment policy and their drop policy due to lack of payment. Not only was the new policy better for students because they could continue their education, but the institution also gained its investment back in two semesters--seven months.

Briggs: Can you share a couple of examples of smaller schools that are doing productive things with BI?

Rouse: Let's talk about Notre Dame College in Cleveland, OH. This isn't Notre Dame University; it's sponsored by the same sisters of Notre Dame, but this college has 900 students. In the last two and a half years, it has tripled its enrollment, primarily because it added collegiate athletics to the list of activities for students. The college's business intelligence solution helps it understand how the change in class size impacted student retention and student grades--and specifically how the academic performance of its athletes compared to the performance of non-athletes. With some changes in intercollegiate athletics based on that data, Notre Dame has gone from 300 or 400 students up to 900 full-time students in two and a half years, a fantastic growth rate.

To give a different example, Cuyahoga Community College, also in Cleveland, has a large number of students, but some of the off-campus locations are very, very tiny, and section management is critical. They look at capacity and fill rates every day; it's a significant planning tool to understand whether or not they need to add new sections. In particular, many students take developmental education courses at Tri-C, so they have to understand the course mix that gives the student the best chance for success. They have a lot of stop-out, dropout. As the economy takes a dive, enrollment goes up. As the economy gets better, students tend to go back to work. The college has to anticipate externally what's happening to the economy in order to anticipate how many students to plan for. Their BI system helps them do that.

Briggs: If I asked you for a couple of rules for BI in higher ed, especially for smaller institutions with limited IT resources, what would you suggest?

Rouse: First, listen for the big problem. Listen for the question that gets asked over and over but never gets answered. That should be the first topic of your business intelligence discussion.

The second rule of thumb is look for the skill set, and don't worry about what title the person holds or what department they're in. People in small schools often wear multiple hats, so don't let a department or a title fool you.

Another thing small schools can do is connect with a large school with graduate programs, to get college work-study students to help with their BI solution. Small schools can leverage graduate programs from larger schools. Notre Dame is 15 minutes from Cleveland State, which has lots of graduate programs. Graduate students in the MBA program and the Computer Science program that have work-study funds can work off-campus at another non-profit organization and the federal government pays 75 percent of their hourly rate--the small school could pay just 25 percent. Graduate students can't do the project on their own, but with some guidance from us as consultants, it's a great BI solution at a great price. Plus, it's a great educational opportunity for those graduate students.