Transformational leadership needed to pursue data-driven ethos

Brendan Aldrich, chief data officer at California State University, isn't just a data governance or data quality leader. He also leads the university system's business intelligence and data warehousing programs. And he's a prime example of how the role of the chief data officer is changing.

According to Gartner's most recent survey of high-ranking data professionals, 85% of respondents said they are defining the organization's data and analytics strategy, tackling responsibilities such as data management and even data science.

Prior to being hired by CSU, Aldrich was the chief data officer at Ivy Tech, where he transformed how Indiana's community college system accessed and interacted with data. There, he helped spearhead data-driven initiatives such as Project Early Success, which uses data to find students at risk of failing their courses early enough to intervene. Now as the chief data officer at the largest four-year education system in the country, he's hoping to pursue a similar mission and develop data-driven initiatives that help students succeed.

A featured speaker at the Real Business Intelligence conference, Aldrich sat down with SearchCIO to talk about what transformational leadership is and how he’s found ways to bring about change.

Editor's note: This interview was edited for brevity and clarity.

At the Real Business Intelligence conference, you're talking about transformational leadership. Can you define what that phrase means and why we need that kind of leadership today?

Brendan Aldrich

Brendan Aldrich: Transformational leadership is about being a game changer -- being that person who, rather than do your job, is going to redesign your job both for the benefit of your organization as well as, potentially, for your industry.

There are three things that I tend to say: We need to make data intuitive, relevant and interactive. As I've grown in my career, I've been in IT 20 some-odd years, I began to realize that for much of the last 30 to 35 years, across all industries, we're using data the same way today that we did 35 years ago, which seems silly because in so many areas our technologies are advancing and evolving.

In my work, especially in the last six years in higher education, what I've been focused on is how do we change the rules of the game? How can we start to leverage advances in technologies to be able to advance our capabilities around information? How do we do things that are better than what we've done for 35 years -- and do them less expensively and with less people to either build or support it?

A lot of that comes down to asking some initial questions: Why am I doing what I'm doing? Is there a better way to do this? How do we start to address some of these challenges with data that we've accepted for 30 years rather than solved?

There's so much to unpack in that response, specifically this idea of transforming how companies use data. You've been a champion of making data available to the masses at Ivy Tech and now at CSU. Why is this so critical?

Aldrich: The three words I mentioned earlier -- intuitive, relevant and interactive. If you can do those three things well, then you can open up data to thousands of employees across your organization for them to use intelligently and accurately to test an idea or prove out a theory, which could revolutionize the way your organization operates.

Your job description, which includes developing a big data strategy and building a data lake, is a pretty tall order. Where do you get started when taking on such an expansive role?

Aldrich: As a chief data officer, I often say that my job comes in two phases. Phase one is getting my arms around what exists and what we're doing today. The tools I need to help the organization capitalize on data typically don't exist when I first arrive. So the first phase of my job is figuring out what we have and what tools we need so that we can start to bring our data together to capitalize on it. That's what I'm doing now.

What is such an interesting challenge here at CSU, because we have 23 independent campuses, is my approach has to not only ensure statewide consistency on a variety of metrics, but it also has to support and enhance local campus diversity -- the ability for individual campuses to iterate on this data, to create new measures and dimensions that maybe aren't in use at the chancellor's office but are critical to a campus.

Once we do that, and that's usually the first year of the role, then I switch to becoming much more consultative. I'll be spending time with different teams -- the registrars, the advisers, the faculty members -- and asking what they need.

At Ivy Tech, that's where some of the initiatives such as Project Early Success came from. With 77% accuracy, Project Early Success predicted which students were likely to fail which courses and why in the first two weeks of the term using behavior-based models.

Let's dig into Project Early Success a little. How did it work and what did you learn?

Aldrich: Project Early Success utilized technology and data to find students who were just starting to struggle early enough that you could intervene with the right piece of advice and change that trajectory.

When we did Project Early Success at Ivy Tech, we captured a lot of information. The first term we did it, we had about 60,000 enrolled students, and we predicted 16,247 of them were likely to fail one or more classes.

During weeks three and four of the term, I helped to coordinate over 100 faculty, staff and administrators to make 20,053 phone calls to those students. We eventually reached 31.5% of them, and we captured information about those conversations so that we could study them later.

In 11 cases, we found the reason the student's behavior changed was because their heat had been turned off. Now, when your heat is turned off, you don't call your college. You might call your parents, you might call your friends, you might sell your television set, but you don't call your school.

As it turned out, Ivy Tech had an emergency funds program to do things like help students get their heat turned back on so that they could focus on studying. And I think a lot of colleges in this country have a whole range of services for students that they don't know about. Being able to use data is sometimes just a matter of helping to find students who need those programs and make that connection between the two.

I think for us here at CSU, as we get our data together and we start putting these platforms into place, that will be one of our first focuses: how we help ensure that our students are able to succeed.

What do you see as the biggest data challenge at CSU and how are you planning to tackle it?

Aldrich: This is not just CSU, it's not even higher education, but I'd say the biggest problem facing companies today when it comes to data is addressing the cult of opinion. There are things that sound logical and reasonable, so we all believe them. When we started to do Project Early Success at Ivy Tech, one of the things we heard over and over again is that students don't want us to call them, that we call them too much, that they hate hearing from us.

So what we did is when we called students, every caller filled out a form about the call. One of the things we asked callers to measure was did the student seem happy to hear from you. Overwhelmingly, responses to the calls were in the neutral to very positive range -- almost 98% of the data captured.

In fact, that very first term we kept hearing over and over, 'I can't believe Ivy Tech called me. I mean, Ivy Tech is such a large organization, but yet they called me to see how my term started.' That's the kind of power that, when you're utilizing data correctly, can make a difference.