The Data Chief | Episode 2

VMWare's Stephen Harris on navigating imperfect data

Stephen Harris



Current EpisodeEP2: VMWare's Stephen Harris on navigating imperfect data

Episode Overview

Today’s guest is Stephen Harris, the CDO of VMware. Stephen has more than 20 years in the data, analytics, and professional services fields. He and Cindi discuss growth, diversity in tech, being a change agent within your organization, bias in AI, data for good, the talent gap, data fluency, and more

Key Takeaways:

  • The importance of understanding the role of data and its potential impact when there is an absence of fact or misinterpretation of data.

  • How Stephen and his team at VMware are using data to navigate the current COVID-19 coronavirus crisis — internally and externally.

  • What a flawed facial recognition demo tells us about AI and the risk of bias at scale.

Key Quotes:

From EDWs to the era of knowledge management to the era of business intelligence, to the era of, now, analytics, advanced analytics, and driving insights is a core component of the strategy of many corporations, I’ve served in many roles. I had an interesting opportunity where my team and my data platform was core to the [NASA Space Shuttle] Columbia accident investigation. So if you remember, STS-107 disintegrated upon reentry. That became a very pivotal moment, not only for me, but for the team that I was leading as a contractor working for the Ohio Aerospace Institute, directly contracting with NASA. And so that really started to shape and inform not only my career path, but also the importance of understanding data, its role, and what the potential impact can be when there is an absence of fact or misinterpretation of data.

I learned the importance of how to properly represent the data, not only to support storytelling, but also to drive things like what is the risk profile? How should leadership think about the context in which we are framing the data to go along with the story? And what is most relevant and important?

I think understanding diversity really requires understanding that it’s more of an equation, right? It’s diversity plus equity plus inclusion are sort of the foundational requirements to getting to a place where this industry and others — tech in particular — are able to adopt, adapt, and transform.

I started to realize that my data literacy was at 90 percent, whereas the other individual’s data literacy and the practices around understanding data, how it transforms through systems and then shows up for the consumer to use it, we’re on two different ends of the spectrum. And so my job was to educate, to inform, and to bring them along the journey, as was it to learn and to listen more than I speak. And I always have to throttle that in order to be effective.

More on culture:

The mentality of us, we, they, as data professionals, is that what gets measured gets done. Well, culture is one of those things that’s very, very intangible and hard to measure. Even though we have mechanisms like employee surveys, you the voice of company X, Y, and Z, so there [are] ways to get certain signals that give us indicators. But again, culture is one of those things that’s difficult to really measure, and you have to sort of figure out what’s most relevant to align what you measure to the sort of core values of the company and keep the pulse there as you drive and execute on whatever the set strategy may be.

Guest Bio:

Stephen Harris has over 20 years of professional services experience, managing large-scale project teams across a variety of industries for multinational companies and federal, state, and local governments with 400+ team members and budgets exceeding $1B.

Stephen joined VMware in 2018. As the company’s CDO, he leads a team committed to evolving VMware’s data and analytic insight capabilities with the needs of the customer taking top priority.

Stephen is also an adjunct professor at San Francisco State University, teaching the next generation about big data analytics, data mining, data architecture, and visualization tools.