Consensus, Compromise, and Persistence: Implementing a Single ERP for 13 Colleges

Julie Ouska is CIO and vice president, Information Technologies, for the Colorado Community College System.

The Colorado Community College System (CCCS) comprises the state's largest system of higher education, serving more than 159,000 students annually at 13 state community colleges. Yet all the administrative processes that faculty, staff, and students require at each location — as well as all those needed at the system office — today are met by a single shared database.

The model is called multi-entity processing (MEP), and, for CCCS, it has transformed 13 instances of enterprise resource software (ERP) into a single installation of Ellucian's Banner, which handles all student, financial aid, human resource, and finance transactions. Updates to accommodate financial aid and legislative changes are implemented only once. MEP lets us unify 13 sets of values without losing each institution's distinguishing characteristics.

A State-Mandated Model

The requirement to be on a single ERP, which was mandated by our state legislature in 2004 via House Bill 1086, has helped us achieve critical goals in cost savings, consistency, and control. At the time of the mandate, our system office staff supported 13 separate versions of the same customized ERP software; maintaining 13 different versions of each administrative module required considerable IT support.

CCCS also faced challenges related to different applications of official procedures. For example, some colleges used grading procedures that included pluses and minuses; others did not. Policies for withdrawals and non-payment also varied, as did the interpretations of new statutes. Although no process was actually incorrect, we needed consistency across the colleges to better analyze data at the system level.

Achieving consistent data and business processes required two parallel efforts: achieving agreement on universal processes and policies, and managing the complex technical build and implementation. The 13 colleges were running older versions of Plus, the SunGard Higher Education (now Ellucian) ERP software. CCCS selected Ellucian's Banner as our new ERP software because it could handle all 13 colleges' daily transactions while on a single database shared by each college. The multi-entity functionality allowed each college's data to be separate, such as course enrollment, student transcripts, and financial transactions. Data such as the course catalog, Chart of Accounts, vendor, and general person data are shared by all institutions. The challenge was rolling all the data into a single database while still maintaining separate data and processes for each institution. CCCS and Ellucian collaborated to develop a single multi-entity database that met the complex needs of data segregation and aggregation.

It took us five years to build this MEP environment and go live with the student, finance, financial aid, and human resources modules. Student and financial aid went live first; finance and human resources/payroll followed two years later. The deployment time was longer than it would likely be today because we had to work with the vendor to create the multi-entity software. Our overall cost for implementing all modules across the system was $26 million. The math is compelling: 13 colleges received the latest ERP software at that time for a cost that was far less than they would have paid individually.

Analyzing Processes to Find Common Ground

Although the technical deployment was complex, aligning 13 colleges on consistent business practices was more so. Because our community colleges serve different demographics, we made it a priority to align business practices to meet system-level needs while also acknowledging each college's uniqueness. For example, Front Range Community College has three urban campuses serving 21,126 students, while Lamar Community College supports 935 students from the rural area surrounding it. Despite their differences, the colleges had to agree on a chart of accounts and key registration and tuition-collection policies, including student loan release dates and the criteria for issuing an incomplete for a course. Fortunately, the colleges had previously worked through a similar process when common course numbering was implemented. Still, it was a major effort — but worth it.

Functional leads from each institution's key departments — registrar, financial aid, human resource, and finance — met every week at the system office and worked diligently to compare, contrast, and ultimately agree on data definitions and processes. The effort and time commitment required was significant. Ellucian consultants led the functional groups through analysis of the business processes, and then determined how Banner could support the proposed methods. As agreements were reached, Banner's screens were configured appropriately, and the consultants trained the lead functional users, who then trained the staff at their individual colleges. Process agreement and training took 18 months.

Because our system comprises only community colleges, we have significant commonality; this approach might not work in more diverse systems, such as those with both community colleges and research institutions. However, if you can address 85 percent of your needs with similar processes and then determine how to handle the differences, this approach is worth considering. Also, the procedure for analyzing business processes would be valuable to any institution or consortia — even if they are not moving to the MEP model. The process required our users to question their traditional ways of working, which helped us identify and eliminate many less-efficient tasks.

Consistent Data for Comparison and Analysis

One of this model's biggest benefits is that it offers a system-wide view of data. This lets us identify issues and trends, such as resident versus nonresident student enrollment, as well as changes and opportunities in our online population.

Financial data is also much richer. Prior to the MEP model, each college had its own chart of accounts. Now, our system-level finance representative can work more efficiently with an individual college to help it assess its expenses and budgets. The consistency in data makes understanding each college's financial picture much clearer; it also eliminates all the conversations about how numbers were reached. Enrollment is counted the same way across the colleges. Before, understanding the CCCS system as a whole required considerable — and often difficult — manual work and manipulation. Said Mark Superka, vice president of finance and administration at CCCS:

"[The MEP model] gives us visibility across all the institutions and consistency of data. We can monitor spending and revenues and see trends in enrollment and funding, and better plan for them. I imagine that almost any system with multiple institutions would find this approach valuable."

Today, we can easily see the total paid to CCCS adjunct faculty in a given year. We can monitor the retention of first-time, full-time students, and we know that all colleges have the same definition for that cohort. Revenues are recorded in their proper categories and consistently coded. We can rely on our data, as well as identify abnormalities and why they exist.

Building the MEP Environment

By far, our MEP model's biggest cost savings comes from the decreased support and maintenance of our administrative systems. One team runs a single instance of Banner for 159,000 students across 13 colleges and the system office far more effectively and less expensively than if each college maintained its own administrative systems.

MEP is supported by 50 IT staff members, who provide all other technical services, including supporting a single WAN, a system-wide Microsoft Exchange, and consolidated VoIP for 11 of the colleges. The staff that supports the Banner environment includes Linux and system administrators, project manager/business analysts for each functional area, and quality assurance analysts and developers. The annual budget for IT staffing and maintenance — which is shared among all the colleges — is $10 million. The colleges maintain relatively small IT staffs onsite to support local systems and provide desktop support.

The MEP environment has baseline multi-entity functionality throughout the Banner modules, greatly reducing our maintenance time. We've also implemented several solutions that we developed or modified to work in the MEP environment, including Banner Advancement, Ellucian Degree Works, Banner Workflow, Ellucian Luminis Platform (our portal), Banner Operational Data Store, Banner Document Management, and Ellucian Mobile. Our collaborative relationship with Ellucian continues as we implement new solutions and the company further refines its MEP implementation across all products. The challenge remains getting the software to achieve a balance between providing individual colleges with data and applications, while also rolling data up to the system view. For example, each college retains its unique code for financial aid; however, we can see financial aid data in aggregate for the system.

In terms of solution selection, we consider solutions only from vendors that can provide a multi-entity interface, which lets us implement and maintain new solutions without significant time on our part. Most vendors can deliver the interface; their bigger hurdle is scalability. Consider these metrics for the transactions executed through a single Banner instance:

• In one month (August), we executed almost 1.5 million student transactions (registration, adds, and drops).
• Our maximum number of transactions on a single day was 142,056.
• We paid out almost $422 million in financial aid in a year, and processed a monthly payroll as high as $27 million.
• Electronic tuition payments through self-service Banner were nearly $15.5 million in one month.

Another important consideration for supporting this complex model is that it requires a high level of IT skill and understanding. It also requires outstanding communication, listening, consensus building, and collaboration skills. In this environment, decision making and even implementing upgrades and process changes take significantly more time; for example, automating Return to Title IV [https://fafsa.ed.gov/privacyR2T4.htm] (R2T4) was a year-long project because each college had slightly different manual and semi-automated processes. Also, the new automated process had to work for both the largest and smallest colleges.

More Functionality, Less Cost

Working as a unified system requires weighing consistency needs against each college's individual priorities, resulting in campus-level gains and losses. Michele Haney is president at Red Rocks Community College, which has an annual enrollment of approximately 14,000 students. She has found that sharing information helps the colleges provide improved services to students who attend multiple institutions:

"Our advisors have a more accurate and complete picture of a student without having to request and wait for a transcript. Knowing their true academic history and performance helps us to help the student. It strengthens our academic advising."

Testimonials to the Colorado Community College System's commitment to student success (3:29 minutes)

However, in keeping with FERPA regulations, only limited student information can be viewed across the colleges. For example, the student's academic record is specific to a single college, yet their General Person information is available to all colleges. In addition to protecting students' privacy, our firewalls and programming make it easy for colleges to communicate directly with their respective students. A college can send an e-mail of interest to all of its astronomy students, with confidence that only those students will receive it. These barriers were created programmatically to eliminate accidental processing.

Students also benefit in that they complete a single common application and a single placement, both of which are made available to all colleges. Further, students in smaller colleges gain access to functionality that their institution could not afford on its own. Front Range Community College developed a student tracking system and pioneered a smart catalog that all colleges now use. As another example, a researcher with the system's online consortium recently created a tool to analyze enrollment and course offerings through CCCOnline. Because each college has consistent data elements, software can be shared and used effectively by all the institutions.

"Without this model, most of the smaller colleges probably wouldn't have access to the functionality we've developed," said Andrew R. Dorsey, president of Front Range Community College. "We are saving money as a system; it clearly makes sense economically to take this approach."

The system also provides operational support for the smaller colleges — such as processing financial aid applications — which gives campus staff more time to work directly with students.

Finally, the individual colleges can share best practices. Before, each college had its own way of managing adjunct faculty. Through a comparison, we determined that two colleges had the best process management practices, so we configured the ERP to incorporate those practices and rolled them out to every campus. We've made similar modifications to model best practices in financial aid, payroll processing, and accounts receivable.

In spite of the commonalities, the colleges have retained their unique identities. When a student visits a college's website, the look, branding, and resources are specific and customized to that college. Students register for courses, make payments, and perform all processes with the college, not the system. And the colleges have some flexibility, such as setting their own semester start and end dates.

Consensus vs. Tailored Functionality

Although there is considerable commonality across CCCS, there are also significant differences — especially between our smaller rural and larger urban campuses. At times, it is difficult to balance and prioritize institution-wide needs with those of functional users.

Case in point: because a large portion of the 30,000 students at Front Range Community College receive financial aid, automating those processes was a top priority for the college. However, prioritization had to be collaborative and representative of all the colleges' needs; as a result, financial aid automation took longer to implement than Front Range had originally planned.

Structured Governance: Achieving Balance and Consensus

Balancing individual college priorities with other system institutions, and with the system as a whole, requires multiple layers of governance.

At the system level, we have a provost and other senior executives who interact with the vice presidents and other leaders at the individual colleges in areas such as instruction, student services, registrar, institutional research, finance, human resources, and financial aid. At this level, members work closely to coordinate the colleges' needs with the system's priorities.

Functional decisions are made by representatives from all 13 colleges. For example, business officers are involved in decisions affecting payroll, while presidents have the final say when a group requests a modification or new functionality that has an associated cost. The presidents also must evaluate any new processes that impact policy.

The costs associated with the new software have been divided among the 13 colleges. However, each college can opt out of new functionality and its costs. For example, several of the larger community colleges recently wanted an optional Banner module that smaller community colleges had no need for; the software's cost was thus divided only among colleges that requested it.

An executive ERP steering committee weighs requests that impact the enterprise IT infrastructure. That committee includes the system provost, system president, system CFO, three college presidents, system legal counsel, and me (as system CIO). As the primary governance committee, we help define the overall strategic direction of technology investments.

Defining the governance process was a significant undertaking. We had to determine at what level decisions would be made, who would make them, and how the groups would work together. We still are refining our process of prioritizing projects in a way that satisfies all the institutions.

"With so many needs and so many different colleges, developing a consensus on priorities takes work," said Dorsey, "but overall, we do a good job. It does take a fair amount of time and energy to manage the process. One of the harder parts is balancing the needs and interests of different functional areas, like financial aid or student records, with overall system strategic priorities."

The discussions and concessions are ongoing, as new software is considered or new regulations are issued. Recent changes in Title IV funding and to ensure that all colleges are consistent in their return of Title IV funds prompted an eight-month discussion about the associated processes among financial aid directors.

"The process of getting everyone to agree is critical," Dorsey asserted. "You have to think through the business processes every time, and then match the software application to the business process. Having conversations about your business processes is a positive step, but you have to allow time for that."

Admittedly, our system moves more slowly than that of an individual college. But once a change is agreed on, the system staff implements and maintains it, and it's reliable. Further, individual colleges are free from having to support it.

Lessons Learned

Through this ongoing process, we've learned some valuable lessons.

1. Working through business processes and coming to agreement will take longer than you think. Our eighteen-month timeframe was ambitious. Two years would have been more realistic; it would have also allowed more time for thoughtful review of decisions before implementing the student and financial aid software.
2. Develop a comprehensive communications plan to address the cultural change. You can expedite the transition by communicating why the change is necessary, what benefits will result, and how long it will likely take.
3. An effort of this magnitude requires leadership at the system level, as well as an organization that's truly committed at all levels. CCCS was fortunate to have steadfast commitment throughout the process.
4. Align your business practices before you train people on the software, rather than doing the two in tandem, and allow adequate time for training.
5. Have enough IT staff to effectively implement the solution.
6. Scalability and stability are critical. If an application cannot scale to handle peak demands, services will be poor and staff and students will be frustrated.

Conclusions

Today, all public institutions are struggling with the need to reduce costs while still providing the best possible administrative and student services. Given these pressures, most consortia have at least discussed shared services. The first and most common model is consolidating hardware to support individual ERP installations at multiple institutions. I've spoken with people at systems that have a consolidated data center supporting 13 separate instances of an ERP. Yes, they share hardware and staff resources, but they still do tasks — including updates — 13 times, which adds up to tremendous time and cost. The same is true if the databases are hosted in the cloud.

The second model has all institutions within the organization on a single ERP instance. The most rudimentary MEP model, it opens the door to efficiencies in IT support and the benefits of sharing applications and hardware for economies of scale. However, in this model, institutions do not standardize their business processes.

We at CCCS believe our MEP environment offers the ultimate level of shared services. Through it, we have gained tremendous savings in our IT costs, and students throughout the system have access to more technology applications at a lower cost. Equally important, the standardization of data elements and processes delivers ongoing operational savings in our business functions and enables effective data analysis across the system, helping us make better decisions in support of our 13 colleges and our students, faculty, and staff.