At the latest GovLab Ideas Lunch, a joint event held by NYU Wagner and NYU Stern, Anita McGahan, the Associate Dean of Research, PhD Director, Professor and Rotman Chair in Management at the Rotman School of Management at the University of Toronto, gave a talk on Unlocking the Big Promise of Big Data. Rather than focusing on the technical components of big data, McGahan instead discussed the wide array of management challenges and opportunities related to the burgeoning field.
The unquestioned starting point for the talk was: Big data is everywhere. Big data is being leveraged in contexts as varied as land use, gun permits and consumer interest in “stretchy jeans.” McGahan discussed, in particular, three evolving means for collecting big data and challenges related to each:
- Evaluating the quality of suggestions;
- Developing criteria for choosing among legitimate alternatives;
- Open Innovation Platforms
- Structuring the ask;
- Identifying true expertise;
- Building complex capabilities for integrating expert knowledge.
- Social Media
Whatever the data source, the belief that big data has the capacity to not only make institutions more effective and efficient is widely held, but McGahan argued that big data can also make institutions fairer. In fact, she made it clear that her current life mission is using these tools that are currently being used mainly to bolster big businesses’ bottom lines to help make the world a better place.
However, to unlock the full potential of big data to benefit the public good and avoid a number of potential pitfalls, we must navigate a number of management challenges.
Managing Big Data Culture Change
McGahan’s central argument revolved around the need for building a culture of management for big data. There will not be meaningful, strategic, evolving success in this area if change and decision-making only comes from the top. She referenced the successful, unified culture change initiative undertaken by Jamie Dimon at Banc One as a potential blueprint to follow, and the failed governance innovation undertaken by five different Kodak CEOs as an example of innovation failing without culture change.
A large teaching hospital’s midwife diagnostic program provides another example of how a well-managed, agile big data system is essential. The hospital’s midwife diagnostic program is a system to train midwives to determine whether or not a pregnancy will have complications. When they took the program to South Sudan, somewhat confusing data emerged from different trials – like diminished performance when teams self-selected their members, which turned out to be a result of lowered feelings of competition and more socializing. McGahan warned that without the management knowhow and culture to draw useful insights from the at-times-unintuitive analysis of a great deal of data, results could prove too confounding to be useful.
Managing Big Data Reuse
The beauty of big data – and also the cause for some of its more troubling implications – is that data can be useful long after its original collection and in different contexts. For instance, McGahan highlighted Google Cloud Platform’s business model of enabling the sale and use of data ex-post, proving that big data should not lie dormant.
The demonstrated ability to use big data in new contexts is especially encouraging when considering public problems. McGahan called for the application of insights drawn by large companies with sophisticated big data operations to public problems, like stopping communicable diseases, where there is not a clear, easy path to big financial rewards.
Similarly, just as the burgeoning field of open data is showing how publicly held data can be leveraged for new uses, McGahan called for more work to figure out the optimal incentives for businesses to open up their data. Managing big data initiatives with the future in mind – even a future built outside of a businesses’ walls – could help to solve public problems. McGahan pointed to the paradigmatic example of Glaxo Kline Smith putting failed HIV research online to ask experts to help figure out the problems with their approach as a strategy to emulate.
McGahan was careful to note, however, that while data can be applied in different contexts, cross-sectoral analysis can prove difficult. Big data is built on correlation and analysis, and when the context shifts, the quality of inference can diminish, making it more difficult to come to actionable insights. Moreover, management must ensure that principles subscribed to for one use of big data are adhered to when the data and insights are reused.
Managing Big Data Implications
McGahan clearly sees a bright future for the public arising from the advent of big data. Nonetheless, a number of challenges and implications still exist. She highlighted, in particular, the privacy implications of the widening use of big data. The mapping all gun permit holders in Westchester County, NY is just one example of how the increasing availability of data and the ability to visualize it in easily comprehensible ways can lead to actions with which many people would feel uncomfortable. Perhaps the biggest challenge for big data going forward will be finding ways to dynamically protect privacy.
Additionally, management challenges stretch from inclusiveness concerns, such as the Burnett, Wisconsin’s vacant land data being used by developers to spot all available marsh land in the area, to fairness and false positive concerns, like proposed initiatives to use big data is to help determine the likelihood of recidivism in juvenile offenders.
Managing Governance Innovation
While the sheer amount of data available today, as well as the advancing technological capability to analyze it, have changed the way businesses and governments make use of data, McGahan argued that challenges related to the public leveraging of big data include many of the familiar management issues of governance innovation, like assuring stability in the delivery of essential services, enabling fair and effective process, negotiating and aligning incentives, narrowing scope for experimentation and expanding scope to scale up subsequently.
Of course, if public big data initiatives are going to find success, a meaningful system for tracking progress toward goals will be essential. McGahan noted that such a system requires attention to qualitative goals and a mechanism for confronting disputes over the legitimacy of measures.