Data-Driven Decision Making

Stir Up Your BI Initiative!

• By Graham Tracey
• 03/01/08

Without all of the ingredients necessary to move your campus culture toward an evidence-centric model, your new business intelligence technology is just software.

A 'culture of evidence': The term gets tossed around mightily in higher education these days. But what does it mean? Is it just another catchphrase, or is there a chance that the concept-if realized through to implementation at colleges and universities across the nation-has the potential to change higher education as we know it?

Simply put, a culture of evidence is the outcome of a deliberate strategy to make decisions through the analysis of data. A culture of evidence cannot be forced and it is not something that is implemented. What's more (and as most higher education leaders know), the culture of an institution does not change overnight; and it certainly does not change without strong leadership and the right tools in place within the organization.

Spearheading External/Internal Accountability

At Oklahoma City Community College (OK), President Paul Sechrist has embarked on just such a journey to change the culture of decision-making at the institution. While OCCC's Office of Institutional Effectiveness had been providing senior staff with regular monitoring reports, Executive Director of Strategic Planning Stu Harvey explains that "Dr. Sechrist has made it a priority for staff and faculty to build on this activity and use better and more timely information to make decisions. Anecdotal decision-making is being discouraged, though it will be with us a little while longer, I'm sure."

Sechrist is not alone is his quest. Many college and university presidents across the US are now insisting that decisions be justified by data and not simply arrived at via hunch, intuition, or best guess. Presidents are heeding the call for increased accountability to their constituents as though their posts depend upon the change-and in some cases, they do. As we all know, colleges and universities have been challenged to become more transparent to the constituencies they serve. The US Department of Education and accrediting agencies, along with students and parents, are demanding more detailed information about the success of past alumni, graduation rates, admission criteria, and other critical data.

Some institutions are not only meeting the challenge, but a few, like Slippery Rock University (PA), are now publishing in-depth data on their websites- highly detailed information regarding admitted class profile, retention, and graduation rates, for instance. (See SRU's "Rock Solid-Institutional Profile: Accountability 2008" site here.)

According to SRU President Robert Smith, "The impetus for [this accountability website] came from Secretary [of Education Margaret] Spellings' Commission on the Future of Higher Education's final report: 'A Test of Leadership: Charting the Future of US Higher Education.' These are useful, highly relevant data that enable the public to make informed decisions about whether or not Slippery Rock University 'fits' education wants and needs, where the institution stands relative to its peers, and how well the university delivers on its promise to provide a 'rock-solid' education."

Importantly, institutions are also sharing more information internally: Faculty and administrative staff can now head to employee intranets and portals for data views that can inform or spur action. Of course, when the public display of information highlights lessthan- stellar results in particular areas, it's not uncommon for faculty and administrators to defensively rationalize the results. As Oklahoma City Community College began to make more information available internally, for instance, some faculty members reacted. Says Harvey, "You have to expect a certain amount of resistance. But," he adds, "I've actually been pleased with how receptive most faculty members have been to the information. It's been my experience that most faculty are passionately committed to the success of their students and want to be part of the solution."

Enabling Institution-wide Data Access

Still, there is an issue of access. In a culture where it is mandatory that decisions be supported by evidence, there is, understandably, an increased demand for access to data. Where once, information access was within the domain of a few "power users" adept at querying the database, a whole new class of users is now demanding that same right of entry, but these users want access without complexity. Enter business intelligence (BI) solutions, designed to provide the tools to extract and "surface" data from operational systems and help decision-makers spot trends and make informed decisions. While the tools have proved essential, in and of themselves they do not constitute a culture of evidence-they can, however, form the foundation upon which that culture can be born. A culture of evidence takes time to develop, and must be nurtured. Business intelligence methods and tools, when properly deployed, will help enable the environment from which this culture can emerge.

In response to their users' demands for increased information access, IT departments have turned to a new generation of BI solutions that now provide the means to optimize data for strategic decision-making. This task can be quite complex in higher education, as data must often be combined from disparate data sources such as the student information system (SIS), learning management system (LMS), and constituent or customer relationship management (CRM) system. BI products from vendors such as Business Objects, Cognos, SAS Institute, and Microsoft aim to solve this problem and provide an enterprise framework for surfacing information to campus information consumers.

BI Implementation: Overcoming Barriers

Although the software from the BI providers is comprehensive in nature, there are many barriers to a successful implementation. As with any enterprisescale implementation there will be issues with people, process, and technology. In fact, in some ways the implementation will remind many of when they first implemented a new SIS or enterprise resource planning (ERP) system. This is because the creation of a shared-information platform will require the institution to revisit previous technology implementation decisions. Business rules embedded in operational systems will require decoding, business processes will need to be mapped, and data structures will beg documentation in a language that can be understood by technology professionals and decisionmakers alike. No easy task!

The biggest mistake, however: to approach the BI implementation from a tools and technology perspective. While there is certainly a great deal of technological know-how required in implementing business intelligence, the bigger problem is understanding the institution's business and how the information may be used. There are three main ingredients for success:

Know Your Ingredients

1) Technology. BI tools are complex. They have elaborate security schemes that may need to be implemented differently than those deployed in other campus operational systems. Then too, it is not uncommon to find that the suite of tools available through a single BI software provider is really comprised of several 'loosely' interfaced tools that are not fully integrated. Clearly, to make a seamless experience for the information consumers, single-sign-on technologies and authentication methods (such as LDAP) must be employed.

While these technologies will help to standardize security and provide fewer passwords for users to remember, there is still the question of the front-end experience for end users. For instance, BI software typically includes a proprietary portal for viewing web-based reports and analytics. The question for IT professionals becomes: Will the BI portal live alongside a previously implemented enterprise portal, or can it be embedded within the enterprise portal framework? These questions must be resolved early and may be dependent upon the features available within the selected BI solution. Technology professionals with an understanding of portal and web technologies will be required to help resolve these user interface concerns.

The technical hurdles are no less complex on the data integration side of the implementation. Often, data will need to be brought together from several different transactional systems. These systems frequently run on different databases and technology platforms and require the expertise of database and system administrators. In addition, given the mix of technologies and expertise required for each of these tasks, it is highly unlikely that a single individual within the IT department will have the background and knowledge to lead the organization through all of the technical issues. Instead, it will be necessary for several individuals from both inside and outside the IT department to become involved in the effort. The CIO may need, for instance, to secure the expertise of database administrators from within the IT department, web experts from within the public relations department, and systems administrators from academic computing.

2) Business expertise. Decision-makers are the ultimate beneficiaries of business intelligence; thus it is vital that these information consumers be integrated into the BI project from the very start. They need to feel ownership of the project and understand that success depends on their active participation and leadership. Just as important, the business or subject-matter expertise for these implementations must come from those who make the decisions. These are not the front-line staff members, but the deans, directors, and vice presidents. Their role in the implementation is to articulate: the goals of the institution; how their organization works to fulfill those goals; and the information needs they have, in order to make informed decisions. While these decision-makers may have associates and assistants to help them execute on their plans for their organizations, the deans, directors, and VPs must not delegate this responsibility to those not directly responsible for the institution's outcomes.

3) Data analysis. This is perhaps the most elusive skill of the three. Data analysts are the people who understand the data structure of the operational systems, yet at the same time they also have an understanding of the business that is being conducted. Rarely does someone hold the title of data analyst on the campus, however. This individual may be the associate registrar, a front-line staff member in the Admissions office, or an institutional researcher. These are the people who query the databases on a daily basis, helping decision-makers find appropriate data in response to their daily questions. They may be running lists of prospects for mailings in the Admissions office, tracking add/drop registration for the Registrar, or preparing the next round of IPEDS (the Integrated Postsecondary Education Data System) reports. These people know how the data are used, where the data live, and how they are or aren't associated with the data in other parts of the system. Data analysts are the translators operating between the decision- makers and the technologists, and their role is critical to a business intelligence project. The CIO as Enabler Even with all of the ingredients in place, it will take strong leadership to overcome the inevitable challenges that arise during the course of a BI implementation. It is absolutely critical that the champion or leader of the initial business intelligence project(s) have the ear of the president's cabinet. Executive leadership will be required to break through the political issues and turf battles that will emerge, as data from sources previously "controlled" by one organization or department are now to be shared in a data warehouse for all to see. If the CIO is not a member of the cabinet, than he or she must make certain to provide regular updates to the cabinet and must ensure that the vice president with overall responsibility for technology is fully engaged in the project.

Ideally, the desire to drive toward a culture of evidence stems from an executive in charge of one of the institution's operational units (e.g., Academic Affairs, Enrollment Management, Finance, etc.), as it is best when the business leaders are fully engaged. Yet this may not always be the case. Many executives understand the need for data-driven decision-making, but believe that it is the CIO's job to go and "figure it out." Frankly, if the CIO finds him or herself in this position, it may be tempting indeed to take this mandate and run with it. But neither business intelligence nor a culture of evidence emerges through an IT project. So, even if you are a CIO charged with leading the business intelligence effort, you must assume the role of enabler, not executor.

Savvy CIOs will create an ally of a campus executive (the provost or vice president of enrollment management, for example), and then foster a sense of urgency in this individual. A smart way to do that: Find a pressing issue involving an elusive question to which that business leader has needed an answer, and use it to demonstrate the need for business intelligence. (For instance, it is one thing to recognize that there is a retention problem because students are dropping out or transferring to other institutions, but it is an entirely different thing to be able to understand the likely causes.) Once the executive sees that business intelligence is not just about what is happening, but can also help to explain why it is happening, that person will be more likely to become a BI champion as well, and help find resources for the project.

Even more effective: Find a way to demonstrate how readily daily decisionmaking would be aligned with that executive's goals for his or her organization, if only the relevant information were widely available to others within the institution. Suppose the advising staff had easy access to detailed information regarding persistence rates of students in developmental classes? If those staffers were able to view such data against several academic and demographic variables, they might begin to devise systematic programs to assist targeted at-risk populations. Such programs would be based on an analysis of the data and would be directly aligned with a goal of the institution: namely, increased retention rates. Help an executive to make these "connections," and all at once, you've got a new executive sponsor sharing outcomes of the project.

Launching the Initiative

No one wakes up one day and decides to implement BI, let alone try to develop a culture of evidence. Even if the environment is right and decisionmakers are prepared to embrace a new culture of transparency and datadriven decision-making, business intelligence must be implemented in stages. After all, a culture of evidence is all about the impact of change on people and processes. For this reason, it is important to start in a focused area and branch out from there.

Such was the case with the Lone Star College System (TX), comprising five community colleges in the greater Houston area. Administrators and technologists employed this focus-by-focus technique when the system launched its Achieving the Dream (AtD) initiative as a catalyst to implement an enterprisewide business intelligence framework, and to help the colleges use data to understand persistence and success rates and increase access to higher education.

Marian Burkhart, newly appointed LSCS director of business intelligence, decided that the AtD program (aimed at helping minority, low-income, and firstgeneration college students succeed) made a good impetus for changing the system's decision-making culture. "Transforming an institution from anecdotal decision-making to one which utilizes a culture of evidence is no easy task, especially within an organization as large as ours. I viewed AtD as an opportunity to achieve two overriding goals: implement a suite of business intelligence products on a district-wide scale, and satisfy the data and reporting needs of the initiative itself," she says. The system was also anticipating a 50 percent increase in enrollment within the ensuing few years and planned to serve almost 60,000 students in its five colleges. Decisions simply had to be made consistently at each of the five institutions, if they were to scale to meet the needs of these new student populations. Ultimately, the consistent use of data or supporting evidence to make decisions would permit a level of autonomy for each of the colleges, while providing a uniform methodology for decision-making.

Burkhart's team of three data analysts (a newly defined role at LSCS) and one database administrator partnered with decision-makers to produce interactive reports that both satisfied the deliverables for AtD, while at the same time provided ways for information consumers to slice and dice the data for further inquiry and exploration. The result: LSCS' business intelligence department deployed the BI framework, underwent education about business intelligence best practices (see "6 BI Implementation Best Practices"), and has since delivered a wide array of information assets to the LSCS community. In the six months since the completion of the initial AtD project, LSCS has produced dozens of information "cubes" (multi-dimensional interactive views of data), hundreds of reports, and has begun work on an executive dashboard to help college and system leaders track progress against key performance indicators. For Lone Star, increased access to information on student persistence in developmental and introductory (or gatekeeper) classes has led to new methods for improving the delivery of these courses and increasing student success.

While it is still too early to track the long-term results of these initiatives, Burkhart reports, "Information flow has improved tremendously throughout the system, decisions are being made more quickly, and a culture of evidence has begun to emerge."

Advice to Live By

It can't be stressed too often: Business intelligence is a strategy, not a thing. The implementation is never complete, nor should it be. You may fully deploy a framework to practice its methods, but it will constantly and necessarily produce change, and therefore introduce the need to continuously implement additional BI-based projects. The institution's executives must continually assist with prioritization and goal setting. Ideally, the information gleaned from each project will lead you to the next.

When asked about the impact of business intelligence at Oklahoma City Community College, Harvey offered this insight: "From where I sit, BI makes three key contributions: First, it allows for the timely dissemination of consistent data to a large number of users. No more second-guessing where the data came from. Second, it should liberate the folks in Research to examine and analyze the data to spot trends. Finally-and I believe this is the most profound impact-the user community will begin to think in different ways and improve what I call the 'decision yield' from the information, by which I mean the number of good, creative decisions and interventions at all levels in the organization."

::WEBEXTRAS ::
Data Mining & Business Intelligence: Open for Business. Case Study: Mining Data To Find Gold. Data-Driven Decision-Making: It's a Catch-Up Game.

-Graham Tracey is director of higher education services at ASR Analytics, a consultancy specializing in business intelligence, predictive modeling, and data mining.