Dr. Barbara Wixom on Data is Everybody’s Business (Scaling Tech Podcast Ep24)

Sep 26, 2023 | General, Tech

In today’s world, data is no longer the purview of data scientists alone. Data is Everybody’s Business, which is also the title of Professor Barb Wixom’s new book from MIT Press. In this conversation, we learn how to monetize data via Improving, Wrapping, and Selling. Barb talks with us about acceptable data use, and shares details on her framework of Capabilities, Initiatives, and Connections, for generating business value from your data assets.

Dr. Barbara Wixom is a Principal Research Scientist at MIT Sloan’s Center for Information Systems Research (CISR). Since 1994, her research has explored how organizations generate business value from data assets. Wixom is a leading academic scholar, publishing in such journals as Information Systems Research; MIT Sloan Management Review; MIS Quarterly; and MIS Quarterly Executive. Prior to MIT CISR, Barb was a tenured faculty member at the University of Virginia (UVA), and is a two-time recipient of the UVA All-University Teaching Award.

Her most recent book is called “Data is Everybody’s Business”, published by MIT Press, and be sure to listen to this episode in order to learn about the different ways that you can monetize data in your business, as you’ll learn, it’s not just about selling data, and that is often the least attractive way to monetize data!

Listen on Spotify
Listen on Apple Podcasts

Watch the video:
Show notes with links to jump ahead are below

Show Notes from Episode 24 – Dr. Barbara Wixom on Data is Everybody’s Business

Timestamp links will open that part of the show in YouTube in a new window

    • Introduction
    • 00:00 Barb’s opening quote: “And so sure, you can sell data, that’s one way to do that, but honestly I find that very unappealing for most organizations because these days, our data can offer a lot of competitive advantage. Our data reflects understanding that is distinct to us as an organization … Why would we give that away? Instead, we should keep that and be using it, monetizing it, in different ways for gain. In addition to selling data sets, we can use data analytics and improve work and improve operations, lower our cost structure, and improve the products that we are selling to increase our sales.”
    • After the opening quote, Arin and David talk about what they’ve learned working with applications with large data sets. David talks about how in data-driven companies, people at all levels of the organization, from data engineer to customer service, all have valuable ideas to contribute to how the company can better use that data. But most organizations don’t do a good job of gathering that input from across the organization. They talk about how in today’s episode we will learn that monetizing data through selling it is only one way for a business to create value from data.
    • Arin introduces Barb with her bio: Dr. Barbara Wixom is a Principal Research Scientist at MIT Sloan’s Center for Information Systems Research (CISR). Since 1994, her research has explored how organizations generate business value from data assets. Wixom is a leading academic scholar, publishing in such journals as Information Systems Research; MIT Sloan Management Review; MIS Quarterly; and MIS Quarterly Executive. Prior to MIT CISR, Barb was a tenured faculty member at the University of Virginia (UVA) where she was one of my professors when I was there in graduate school. Barb is a two-time recipient of the UVA All-University Teaching Award.

Her most recent book is called “Data is Everybody’s Business”, published by MIT Press, and that will be the primary subject of our conversation today.

  • Why is Data Everybody’s Business?
  • 05:20 Arin starts by asking Barb about why is data everyone’s business? Barb talks about how over her career as a professor and researcher she has always felt that everyone needs to understand data, and she’s always emphasized that in her classrooms. In the past however, it was more obvious which pockets of an organization dealt with the data and you could almost see the things happening with that data.
  • Today is different, where Barb says that “organizations need everybody across the enterprise to partake in data. And that could mean helping to create data assets, helping to understand and manage them, exploiting and using data assets for the good of the company’s mission or to produce financial returns and on and on and on.” So the organiation needs to level set data knowledge for everyone.
  • Ways to Monetize Data
  • 06:45 Barb talks about how data monetization is a much broader topic than just selling customer data. It’s a very narrow perspective to think that selling data is the only wya to monetize it. David and Barb talk about data as a resource, and that just like having oil in your backyard, merely having a resource does not automatically make you rich. You have to follow a process to monetize that resource.
  • Barb explains that you can monetize something in good ways and bad ways, and if you’re not careful you can definitely monetize data in unethical ways. Barb speaks about how “Our data reflects understanding that is distinct to us as an organization. Why would we give that away, right? Instead, we should keep that and we should be using it again, monetizing it in different ways for gain.
  • “And so in addition to selling data sets, we can use data analytics and we can improve work. We can improve operations, lower our cost structure, improve the products that we’re selling to increase our sales. So we can improve with data.” Barb also talks about an industry case study of data monetization in healthcare, and how data analysis was used to lower prescription prices.
  • Responsible Data Ownership
  • 12:55 David asks Barb about the concept of data ownership, and points to a growing awareness that people and businesses have about another organization profiting off of data that they created. Who owns that data? Barb says that this concern most often comes up with personal data (like browsing or shopping history), but that we should not only look at that sort of data when considering monetization. If an organization is using data they have about their customers, and using that data to benefit their customers, that is a good thing. Barb gives the example of a bank using customer data to determine where to locate their branches. That is an example of Improving with data.
  • Giving an alternate example, Barb talks about how if a company does not have sufficient access to customer data, that can actually be a problem and lead to different types of customer complaints. She talks about a bank that decided to have a completely opt-in policy on customer data collection and use, but this lead to the bank sending out loan solicitations to customers who already had loans with the bank. By having so little data, they were not exercising responsible use of the data they had access to and customers complained about the unneccesary loan solicitations.
  • Data Ownership vs Right to Monetize
  • 18:00 To further expand on the idea of responsible data use, Barb talks about patient data. A patient’s medical history is clearly private information of the patient and we consider it very sensitive data, which is good. However, we also want to encourage collaboration in healthcare in order to achieve better patient outcomes. If the patient data is transformed and aggregated in a way that anonymizes it, then it can still be used for the typical data monetization techniques of Improving, Wrapping, or Selling.
  • Barb talks about an example of an Anthem group taking a year to de-identify data and inserting synthetic data where needed, that allowed them to still conduct research on that data and maintain its predictive power, while also maintaining patient privacy. This has allowed that organization to create AI models being used for disease prevention.
  • David brings up a positive example, a Google model that uses Iris scan data for public benefit to detect disease in the eye and how it is able to do this on a scale that doctors cannot by themselves, and is a powerful example of how these can be an aid.
  • Employee Consent and Human Dignity
  • 21:45 Barb talks about a case study with GE which is an example of an employer getting human dignity and AI right. GE already had thousands of employees helping to keep nuclear power plants safe, and instead of reducing headcount, they wanted to maintain the same headcount but make those employees more efficient in their work to help improve overall safety and save human lives. To do this, an AI model was trained on documentation in order to help employees with the security processes.
  • This is an example of doing it right, because that model is being trained on human knowledge and human expertise. This is one area of Barb’s research, around questions of “what is the right of the employee who is sharing their very special expertise, potentially giving up their advantage for the good of the company? What is the company’s responsibility in terms of how to treat that employee, how to communicate to that employee, how to gain consent that their information is collected and used, those types of exactly. The point is consent, rational consent is given in the right context.”
  • A message from our sponsor: WebRTC.ventures
  • 27:05 Building custom WebRTC video applications is hard, but your go live doesn’t have to be stressful. WebRTC.ventures is here to help you build, integrate, assess, test, and manage your live video application for web or mobile! Show notes continue below the ad
  • Capabilities, Initiatives, and Connections
  • 27:52 Professor Wixom talks about a framework laid out in her book to help simplify and create a common language in an organization about data: Capabilities, Initiatives, and Connections. She discusses different Capabilities you need. These capabilities include things like Data Management, and having a data platform that helps you serve up and understand data assets.
  • Organizations also need customer understanding, to understand the customer data that they have as well as what may be available in the wider marketplace. Organizations also need to understand what is Acceptable Data Use, because it’s important to be both ethical and stay compliant with the law. This also includes understanding who owns the data that you are using.
  • Once you have your capabilities, you can define Initiatives to use your data. Barb had already discussed the ways you can use data: Improving, Wrapping and Selling.
  • The final part of the model is Connections. The connections inside your company need to inspire innovation with the data, and having cross functional teams is an important part of this. Teams need to be built up of people from across different functionalities like digital transformation, data analytics, compliance, business, etc. The other key part of Connections is to be able to scale this knowlege across the organization, and make sure that it is not just kept in data silos.
  • Using Data to Improve both revenue and costs
  • 36:00 Barb talks about applications of this model in call centers, where AI models were used to help contact center agents reduce call times from 20 minutes to 2-3 minutes by providing answers more efficiently. Arin remarks how this is an area of great interest to our team at WebRTC.ventures and that we work with frameworks that help to incorporate large language models into the contact center.
  • Arin asks Barb to talk more about how data can be used to increase efficiency. Barb notes that too often organizations really dismiss the use of AI and data or data analytics for internal types of operational changes, but about 51% of returns of financial returns come through improving the nature of the organization’s work using data analytics. So it should not be ignored!
  • Improving can be about revenue increases or cost reductions. Barb talks about a bank where they
    started to apply AI and more sophisticated data science to how they were doing bank distribution projects. By introducing these new data analytics opportunities, the first year they saved $35 million above what they were already achieving through their prior activities. She also discusses a retailer that she worked with who was able to use data analytics in order to better optimize when they were doing markdowns on their clothing. And it turns out they were marking down their clothing too soon. They found out if they could just wait another week before discounting, in certain cases that it would be about a $20 million lift.
  • Wrapping with Cost Reduction
  • 40:55 Barb next gives an example of using Data Wrapping to reduce costs and improve customer satisfaction. Capital One was using data wrapping to add features to the credit card transaction statement to help cardholders feel more comfortable that their transactions were not fraudulent. This includeded information about the transaction like a geospatial map and merchant logo. This helped cusotmers to more easily identify a transaction. This initiative reduced the number of customer calls to their contact center about fraudulent charges and so reduced costs. Interestingly it also led to increased revenue to Capital One because their customers were more confident using that card in more places.
  • Measuring the Value of Data Initiatives
  • 42:55 Arin asks about how you measure the value of a data initiative when there are probably multiple variables at play. A cost reduction initiative may be going on at the same time as another increased revenue initiative or a sales campaign. Barb talks about the importance of identifying measurable metrics for each initiative, and carefully measuring those before, during, and after the implementation of a data initiative.
  • The more granular your metric, the better, Barb explains. “In the case of information businesses, if things are such that you just really can’t disentangle what we did with data versus all the other things that went down, then you have to take it at a higher level and be comfortable that data analytics played a significant role in the benefits that are being generated.” Different organizations also have different appetites for measurement, so the more instrumented your organization already is, the easier it will be to measure the value of your data initiatives.
  • Communicating and Selling a Data Initiative
  • 47:35 Arin asks a final question about how engineering leaders can best communicate data initiatives, especially if they are not a data scientist themselves. When that engineering leader comes up with an idea for data monetization of any kind in their organization, How should they sell that to executives within the organization? What do they need to consider before making that internal proposal?
  • Barb notes that engineering managers without a lot of data science background are a great example of who she hopes connects with her book, since it is geared towards helping everyone in the organization understand how to use and communicate data initiatives. The most important thing to do when communicating a data initiative is to understand the problem it is solving and the value it creates. Then build a cross functional team around that who can help analyze it further and create a business case. Barb notes that “it’s not about selling the data analytics. It’s about selling a solution to a problem that probably people are aware is a problem.”
  • Experimentation and learning can also be applied to create a hint of the potential value to be created, since it’s often hard to predict exactly what the data analysis will find. Those small experimental victories can also help make the business case to executive management.
  • Conclusion
  • 50:25 Professor Barbara Wixom’s book is “Data is Everybody’s Business” from MIT Press, and there is a link in the show notes below to buy a copy. You can also follow her on LinkedIn to learn more about her work. In the course of our interview, Barb talked about how this book is really a culmination of her research over the last 29 years since she completed her PhD in 1994 on how organizations create value from their data. She notes that the book is like her “legacy of what I’ve learned over that time and what I believe people need to know and I hope in a simple way so that they can understand it and be a bit more successful with their data.” So check it out today to learn more about why data truly is Everybody’s Business!

 

Links from Episode 24 – Dr. Barbara Wixom on Data is Everybody’s Business