How do companies derive value from data? After studying this question for 25 years, Barbara Wixom, a principal research scientist at the MIT Center for Information Systems Research, has boiled the answer down to three possibilities.
First, companies can use data to improve their processes. This is a very old practice. Second, companies can sell data. This method, too, has a rich history. Finally, companies can “wrap” their data around products and services. A tractor may be “wrapped” with a dashboard to monitor operational performance; a bank account may be “wrapped” with a budgeting tool; a cab trip may be “wrapped” with a fare estimator. These wraps can create indirect value by, say, increasing customer retention and wallet share or boosting customer satisfaction.
“But, to be honest, we know very little about data wrapping,” Wixom said — an acute problem given the ubiquity of data. More than ever, managers need best practices for integrating their stockpiles of information into new product features. “If we’re really serious about analytics features and experiences, I want people to know how to nail it,” Wixom continued. “Let’s make sure you’re not just adding cost to your product.”
In a recent report built around surveys of more than 500 companies, Wixom, along with researchers Killian Farrell – a student in MIT Sloan’s Master of Business Analytics program – and Ronny Schüritz, outlines the fundamentals of generating business value from data wrapping. Their findings can be summarized in three steps.
Design for the four “As”
Effective data wraps do four things simultaneously. They:
- Anticipate by intuiting customer needs. Anticipatory wraps offer predictive and proactive features.
- Advise through the use of evidence-based decision making. Wraps with advisory features provide data and insights that inform a customer’s decisions.
- Adapt by meeting customer needs in a tailored way across different environments and contexts.
- Act, which means that the wrap performs an action to benefit the customer. Wraps that act are integrated into customer processes or behaviors, or they trigger behavior automatically on the customer’s behalf. For example, a bank app that automatically transfers funds to help a customer avoid overdraft fees.
Wixom highlighted a 2017 Cochlear behind-the-ear sound processor as a product that effectively embodies all four of these characteristics. The processor both anticipates that hearing needs shift as people change environments through the day and adapts to new contexts, such as to a crowded street corner or quiet room. The feature advises the end-user of optimal settings through an app; it acts by automatically adjusting its settings to deliver the best hearing for the conditions.
It sounds obvious, but “companies have got to measure the value that a feature brings to their customers and their firm,” Wixom said. “I don’t know if the technology is overwhelming, or what, but sometimes companies think of these features as magic that they add on to products with the hope that something simply happens.”
Measurement is a twofold process. First, because data wrapping creates value indirectly, this should be captured with a mix of techniques like tracking customer usage, A/B testing, or controlled experiments and surveys. Second, companies have to pinpoint the source and magnitude of the value that they’re capturing: how much and in what way does this help customers?
This latter point is essential. Wixom described a scenario she witnessed in which a company that dealt with information was selling data to one company that received marginal value on each purchase — something on the order of 25 cents. In return, that company paid the data aggregator a fraction of a cent. The exact same dataset, when sold to a financial services company, was being used to make multimillion-dollar decisions, so the data aggregator could charge $100,000.
“In one case they earned less than a penny, in the other $100,000, and that difference was all about how the data was being used and the value the customer derived,” she said. “You first have to understand how the customer is saving or making money and on what scale, and then go back to figure out how your organization is going to capture some of that.”
Double down on data science
Historically, a wall has existed between product managers and data analytics teams. When managers needed data they put in a formal request; and these requests were often in the interest of improving the data that was already being captured.
“With wrapping, all of a sudden these teams need to be co-located, as the data analytics folks essentially advise the product people on what kind of data exists and what can be done with it,” Wixom said. This co-location also helps product managers get a better sense of their customers. “Either that, or the product team becomes so savvy in its own right that it builds a data analytics team that flows back into enterprise.”
This partnering is an increasingly powerful source of product differentiation. In 2015, for instance, Schindler Elevator Corporation announced it would monitor its fleet of elevators in real-time and use the information to predict maintenance needs. The company was adopting a preventive rather than reactive solution to elevator failure. “So an elevator — an elevator — is now being bundled with that tech and all of a sudden people start saying, ‘Hey, I want a Schindler because it comes with a dashboard,’” Wixom said. Like the Cochlear, this innovation exemplifies the value of data wrapping. “It’s not just about slapping on analytics,” she said. “It’s about a product owner who really understands her market and her customer needs such that she can develop these features that save or make money for someone else.”