Streamlining IT Operations (and More) with Data

These two institutions are learning how to use data to reduce the manual effort required in all kinds of activities — from expediting help requests to counting actual software usage to improving student communications.

• By Dian Schaffhauser
• 06/27/19

While a lot of attention on the use of data is being directed at improving the learning experience, colleges and universities have long known that it can help with IT operations too. Now data analytics is playing a role in those activities to expedite troubleshooting, justify IT investments and make for more seamless workflows.

At Clemson University (SC), for example, while the IT organization had some level of visibility into network operations (after all, monitoring tools have been around since the 1980s), when problems would surface, the institution's networking, IT operations and storage teams would have to expend a lot of manual effort sorting out what was causing the issue or the outage.

According to Nitin Madhok, director of Business Intelligence & Advanced Data Analytics at Clemson, a more sophisticated approach to the use of data is helping IT improve the mean time to fix or recover, "which means happier students, happier staff and happier executives."

Splunk, one tool at the heart of the effort, has been in use for a while at Clemson. The basic idea behind the software is to collect, index and correlate data in real time into a searchable repository, from which users can obtain alerts, reports and dashboards. Clemson's networking administrators have long used the program for capturing logs of data from switches and routers; operations monitors performance with it; and security grabs information on audit activity and authentication, among other applications.

What's changed, however, is that rather than simply setting up Splunk to notify IT when something "concerning occurs," now Madhok and his analytics team are finding ways to be more proactive in their use of the tool.

When maintenance needs to take place on identity management/authentication infrastructure or some other key system, IT can select a time that's least disruptive to users, "based on data from what times people are logging in and when they're more active [or] less active — and that changes across semesters," he explained. While one might guess that Friday night would be a good window for tackling those kinds of projects, "there are certain departments, courses [and] colleges that also schedule assignment deadlines for Friday nights. So, we have to factor those things in and look at the data that's going into Splunk and come up with the time [when] we can actually do the maintenance."

The use of data analytics is also speeding up issue resolution for the support center. Previously, the center lacked a "lot of visibility" into IT operations, said Madhok. Members of that unit have since been given a level of access that allows them to see how the systems are doing; they have built themselves a lot of dashboards in Splunk and are using it "almost on a day-to-day basis." When a call comes in or a student or faculty member arrives on site with a problematic device, real-time access to operations data allows help desk staff to quickly narrow down what the issue is by flagging whether it's on the server side or the client side. One dashboard panel, for instance, displays devices that are repeatedly doing failed authentication attempts. "That usually happens," Madhok said, "when someone changes/resets their password but doesn't update their password across all devices, so that causes them to be locked out."

Such visibility was made possible with the assistance of a Splunk Admin and Security team that helped develop an "access governance framework" for university data going into Splunk. Now, based on the roles people have or the teams they belong to, they'll "get access to the data that they should and are already approved to have access to." This framework is actively being worked on and is being refined with the supervision of a unit called the Clemson Analytics Team (CAT), which works with business analysts, data managers, data stewards and key stakeholders to tackle questions.

Now, all analytic needs "are pretty much funneled through this newly formed team," which collects about two terabytes of data in Splunk per day, Madhok reported. And the toolbox has other resources than just Splunk, he emphasized. Tableau is in use, as are other utilities — whatever "is sustainable, cost-effective and best to do the job."

Where IT and Business Operations Intersect

IT isn't the only direct beneficiary of Clemson's new CAT operation. Data analytics is also coming in handy for justifying software purchases, which has application in other divisions too.

The university adopted Instructure Canvas as its learning management system in 2017, which already includes a lot of "smaller analytics tools" for use by an instructor or student or administrator, said Associate Director for Learning Technology Matthew Briggs. The LMS also produces reports of archived data, which were "just sitting there," he noted. "We had the feature turned on, but we didn't really have a way to ingest all that data and keep it somewhere and then also read it after it was moved."

Madhok helped Briggs and his department use Splunk to set up some new dashboards that could tap into that archived data, mix it with other data sources and give them views into information they'd never had before.
One obvious quick win is to grab information from the LMS on what third-party integrations or tools are in use. For instance, when Briggs' department faced the prospect of selecting a new online testing tool, first it wanted to find out how much usage the old tool was getting. "And we couldn't really get a firm number from either the vendor or by our own digging," Briggs recalled. "The dashboard Nitin produced for us was able to look long term and actually pull an accurate number of how many faculty enabled that tool in their courses." The conclusion: "This is a heavily utilized tool and we should take our time in picking a new one. There really is a use case here that we need to investigate."

The same type of setup is helping Briggs respond to deans and others who want to understand use of the new LMS among instructors, by department or college.

Data Usage in a Smaller School

While Clemson has been able to plow heavy-duty expertise and tools into its use of data analytics, not all schools are big enough to have the same motherlode of resources. That doesn't mean they can't take advantage of the data they have on hand to improve operations too.

At The King's College (NY), a 520-student institution in New York City, Stephanie Brewster, associate director of admissions and new student financial services (and a graduate of the school), has found numerous places where her department relied too often on manual processes that belied easy scaling.

As one example, the office used a tedious combination of Outlook and mail merge to send communications to prospective students. Financial aid, which was awarded by the admissions team, was managed in an Excel spreadsheet. It had "very complicated formulas in there," said Brewster, "but it created all kinds of other problems because it could never be quite as sophisticated as we wanted."

Two things happened to set King's on its path to modernization. The first was adoption of Salesforce for constituent relationship management. The use of Salesforce Marketing Cloud has enabled admissions to set up "integrated automated journeys" that eliminate the manual e-mailing steps that were so painful in the past, said Brewster. Its use also allows her to set up and send targeted messages by each student type or cohort, which "allows us to communicate much better to our students without taking a lot of time from a staff member."

The second driver was the creation of a technology governance committee where users from the different departments come together to make campuswide IT decisions, including those involving data. One ongoing project undertaken by that committee is mapping out the data journey followed by students from the moment they first hear about the college during the admissions process all the way to alumni giving. The committee has also developed metrics by which to rate new data projects, to "prioritize the ones that are most important to us."

But Salesforce in itself isn't the sole tool helping King's find success with its data work. The institution also needed something that would simplify integration of information maintained by the college's multiple systems. A biggie: pulling student data from the school's current student information system — CAMS from Unit4 (previously, Three Rivers) — into Salesforce.

Brewster turned to Jitterbit to expedite the Salesforce integration work, which allows her to set up connections between the two applications. Previously, she recalled, a job of getting the student data out of Marketing Cloud and moving it into the SIS "took hundreds of hours." Now it takes "less than five minutes."

Additionally, with the use of SQL scripts she wrote herself, Brewster has facilitated a two-way information trade between CAMS and Salesforce. "I didn't want this to just be a one-way connection," she insisted. "It's no fun when a student has already talked to someone in admissions about something but that data didn't get passed into the SIS, so that when they go talk to financial services, they're looking at something completely different, which means they have to retell their story. It causes a lot of frustration. We want to be as responsive as possible to [student] needs and give them the best experience we can."

An advantage of automating such connections is that it builds in "redundancies," Brewster pointed out. "Instead of one person having to be very technical to understand how to do a process, now Jitterbit can do the process."

There have been bumps along the way. Frequently, the various applications that Brewster is working with don't have the same "space" for the same information. "I may have a whole bunch of contact information in one system, but the other system only needs an e-mail address," she explained. "I have to figure out how to connect the systems so that they have the information that they need and not lose anything in the process." The tech committee referees those decisions to decide either "which system wins" or whether the difference really matters.

Also, there's "no one person" who can answer questions about the different systems. That's where the data integration work undertaken by the technical committee will be useful — "to close the loop and [give us] a comprehensive idea of what's going on on campus so we can make better strategic decisions based on the most accurate information."

The Payoff Continues

Both The King's College and Clemson University are ambitious in their plans for future use of data to streamline operations.

Clemson, for one, is considering the use of badge swipe data for a possible project involving its role in an emergency situation on campus. "If there's a fire in the building, knowing who's in the building, how many people are connected and things like that would be useful to first responders," suggested Madhok, adding that the institution is still working out details with stakeholders on that.

But once a school gets "into the integration process and streamlining, you realize just how many more opportunities you have — and more things keep coming to mind," said King's Brewster. "It's like, oh, I can do this next or I can automate that process. It's been hard for me to keep myself from jumping into 100 different projects that I could do."