Background Circle Background Circle

Data Can Define Company Culture: Dr. Tom Redman’s Insights on Engagement

After four decades in data, Redman emphasizes putting “regular people” on the frontlines of data quality management who can support the growth of an embedded data culture that thrives. 

DATA LEADERS NETWORK: You’ve rightfully asserted that everyone in an organization is equal parts Data Customer and Data Creator. What practical efforts are most effective for motivating change in data-quality management: incentivized programs, software tools, or something else?

REDMAN: First, simply point out to people that, whether they’ve recognized it or not, they are Data Creators and Data Customers. Most say to me (or themselves), “I never thought about it that way.” Second, explain what you want them to do and how to do it. I’ve never had to do much beyond this. Bear in mind, these roles are far better than correcting errors so you can get through your day. Some people, certainly not all, find these roles transformative.

 

DATA LEADERS NETWORK: How would you define an industry-agnostic mission statement for data quality?

REDMAN: “As much as anything, the ways we treat data define our culture. We embrace our roles as data creators and data customers, and work together to improve our processes, products, decisions, and job satisfaction. Today and forever.” 

 

DATA LEADERS NETWORK: If data is a company’s greatest asset, it needs to become a value of the org. Every function should have a priority to serve this value. That said, what’s the best way for functional experts (e.g., product managers, marketers, operations, etc) and data experts to collaborate to support a data-driven culture? Is there an ideal org structure or collaboration model? How do these roles work together most effectively?

REDMAN: I cover this topic in my most recent book People and Data. Some key features related to this question include:

  • Regular people, front and center
  • Regular people, significantly supported by the data team 
  • Embedded data managers as the tip of the spear
  • Solid organizational collaboration to promote “data is a team sport” concept
  • Senior leadership roles designed to advance this culture

 

DATA LEADERS NETWORK: Many companies have gotten by without leveraging good data. What are some examples of large companies who have revolutionized their thinking and implementation around data, coming from the old world of thinking and into the new? What were the results? 

REDMAN: Yes, almost all get by with bad data. Before I answer, let me object a bit to the phrasing “coming from the old way of thinking and into the new.” I read it a bit like, “Cmon, gramps, get with it.” It’s hard for one person to change their mind nevermind a whole company. Hence the importance of provocateurs, as I mentioned in my prior article. Exacerbating this, in many cases, the “old ways” served the organization well for a very long time. Tossing them out willy-nilly is not smart. (Further on this: From time to time one hears, “Don’t be a dinosaur.” But dinosaurs ruled the Earth for 65 million years. Give them some respect.) 

Now to answer the question regarding large companies, or at least big portions of enormous ones. Bear in mind, my work with clients is subject to non-disclosure. I work very hard to get them to let me put their stories and results in the public domain. A few do, most do not. So, here, I’ll simply cite those that come to mind and provide links: AT&T, Chevron, Gulf Bank, and Shell. There’s also an article that summarizes results, across the board. 

 

DATA LEADERS NETWORK: What companies have proven able to develop and maintain a data culture, and is a pattern starting to emerge that indicates what it takes to succeed? For instance, are there existing titles that play a critical role or do specific steps need to be in place?

REDMAN: This is an interesting question. First, let me be clear that I have never once advised a client to start work with “Let’s build a data culture.” We’ve started out trying to solve business problems. Maybe expenses were too high, perhaps people didn’t trust the data, and so forth. People had to take on new roles and learn new skills to make progress. As they begin to experience success, their confidence grows. As more gain experience in taking on new roles, organizational confidence grows. Then one day, people say, “Wow, the culture around here is different.” So, from my perspective, “culture” is a lagging indicator. People do something different, internalize it, work with each other in different ways, and the culture changes. 

Said differently, “Culture change is won through deeds, not words.”

 

DATA LEADERS NETWORK: On average, how long does it take for a business to develop a fully embedded data culture?

REDMAN: Building on my answer to an earlier question: The culture changes as increasing numbers of people participate in some sort of data quality project. At some point, enough have done so that “data” becomes a dominant theme in the overall culture. So — a long time!

One thing that can help is employee churn. It helps when those fighting the change leave the company. New people generally arrive with more open minds and less baggage. 

 

DATA LEADERS NETWORK: Data initiatives can be a tough sell. What messaging has proven most effective to motivate managers who aren’t techies to embrace data quality initiatives? What about executives who control budgets?

REDMAN: Gosh, if I had a good answer to this I would be scandalously rich. Here is what we try to do. First, we make clear that data and tech are different sorts of assets. Co-mingling the two almost certainly means you’re managing both poorly. Interestingly, I find that almost everyone gets this point immediately, once it is pointed out! But they don’t think of it on their own, so it builds some trust.

Next, we try to “engage both the heart and the head.” Engaging the head is about the business case. Engaging the heart is about showing how the data program will advance things they value. For a hospital, how you’ll be able to provide better patient care; for a financial institution, how they’ll gain market share; today, how you need data for AI. Finally, we try to break up the work into a sequence of simple, doable steps. No miracles required. 

 

DATA LEADERS NETWORK: What are the typical reasons for businesses to backslide on their forward progress with data-quality management, cleanups, and eliminating errors?

REDMAN: This is an interesting and important question … and one that I don’t have a good answer for. One important observation, however: The reason culture is so important is that it makes it much more difficult to backslide. 

 

DATA LEADERS NETWORK: Are there specific cost and profit centers that get hit the hardest by poor-quality data?

REDMAN: Myself and others have looked, as best we can, to answer this question. So far, I haven’t detected anything significant in terms of basic quality levels. But I am most concerned about AI, because the leverage is so great. A bad algorithm can do horrific damage.

 

DATA LEADERS NETWORK: What’s the most successful way you’ve seen for data to find its way onto the balance sheet or income statement to quantify results?

REDMAN: I’ve not seen this happen. And if I understand correctly, GAAP standards stand in the way, at least in the US. I think it’s a great idea and Doug Laney is working on ways around it. 

At the same time, the more fundamental question is “what will it take for companies to start thinking of their data as assets, whether they explicitly put them on their balance sheets or not. To illustrate, most companies recognize their brand and brand recognition as an asset. But they don’t have a line item on the balance sheet. 

 

DATA LEADERS NETWORK: What are the most compelling discussions happening around ethical AI right now?

REDMAN: The range of issues around AI is stunning, from IP concerns to losing control, and your data to violating privacy rights. I certainly think that training a model with poor quality data is unethical. Similarly, feeding an already-trained model bad data for use and/or re-training is unethical. So, too, is using data of unknown quality because you were just too arrogant or lazy to look. 

Join the conversation on LinkedIn.