Big data is something that everyone in business needs to know about. Below, Tom Davenport discusses how organisations can harness the power of big data and suggests that the stand-alone era for big data and analytics is approaching the end of its useful life.
Several years ago, I developed the DELTA (data, enterprise, leadership, targets, and analysts) model for how to build analytical capabilities within an organization. It is also the basis for an assessment tool used by the International Institute for Analytics (an organization I cofounded several years ago). It may be useful to contrast the five factors in the model across big data and traditional analytics.
Enterprise Orientation for Big Data
In traditional analytics, it’s important to take an enterprise focus—to share data, technology, and people across the organization to achieve your analytical objectives. But for the early adopters of big data—primarily start-ups and online firms—this wasn’t much of an issue. People were anxious simply to get something going, and how it related to other big data and analytics initiatives was not much of a concern.
However, there is one factor that is leading to more enterprise coordination: the integration of big data and traditional analytics in large organizations. I interviewed with twenty large organizations pursuing big data initiatives, and not a single one was dealing with big data separately from traditional analytics.[ms-protect-content id=”9932″]
But despite the fact that big data and traditional analytics are often combined in large companies, a high level of coordination of big data projects across the enterprise is somewhat unusual at this point. In the future, perhaps, we will see companies pursuing multiple big data initiatives across different functions and units, and they will feel a greater need to coordinate those initiatives, as they have done with traditional analytics programs. At the moment, however, this is not a primary concern.
Leadership for Big Data
One key leadership trait for big data seems to be a willingness to sponsor experimental activity with data on a large scale. Big data, at least today, requires some educated faith.
There are some leaders who are willing to venture into big data on faith, however. At LinkedIn, for example, cofounder Reid Hoffman had also been a founder of PayPal, and knew there were substantial opportunities from exploiting online transaction data. He encouraged the data scientists he hired not only to try to develop new products and services, but also to contact him directly if their ideas got stuck in the process or the hierarchy.
That’s exactly what Jonathan Goldman, a data scientist at LinkedIn whom Hoffman had helped recruit, did when he had an idea for a new application that became People You May Know (PYMK). The application recommends people you may want to network with who have background attributes in common with you. Goldman created an early prototype of PYMK, but had difficulty getting the product engineering organization to incorporate it into the LinkedIn site—or even to try it.
After Goldman approached Hoffman with his problem, Hoffman allowed him to create a test ad on the LinkedIn site. The click-through rate on those ads was the highest ever seen.
LinkedIn’s top managers quickly made PYMK a standard feature. That’s when things really took off. Thanks to this one feature, LinkedIn’s growth trajectory shifted significantly upward; PYMK is credited with bringing in several million new users. It wouldn’t have happened without Goldman’s idea—and Hoffman’s support of it.
Leaders of big data–intensive organizations also need some degree of patience. It may even be necessary to keep data around for multiple years before its value is known. Jeff Bezos of Amazon is known for saying, “We never throw away data,” simply because it is difficult to know when it may become important for a product or service offering down the road.
Leadership of big data firms may also require some new senior management roles. There are no examples—to my knowledge, anyway—of “Senior Vice Presidents of Big Data,” but there are some roles that include that function. Take, for example, Nora Denzel, who was the senior vice president not only of marketing, but also of big data and social design at Intuit (and big data actually comes first—her official title there was Senior Vice President of Big Data, Social Design and Marketing). There is a logic to combining these roles; at Intuit, big data is used to improve the website, build customer loyalty, and improve customer satisfaction—all marketing objectives.
There are also new senior management roles at other firms involving the combination of big data and analytics. The insurance giant AIG, for example, brought in long-term analytics leader Murli Buluswar to be chief science officer. He commented in an interview: “From the beginning of our science function at AIG, our focus was on both traditional analytics and big data. We make use of structured and unstructured data, open-source and traditional analytics tools. We’re working on traditional insurance analytics issues like pricing optimization, and some exotic big data problems in collaboration with MIT. It was and will continue to be an integrated approach.”
We’re already beginning to see more roles of this type, with a variety of specific titles. If you are really serious about analytics— not just the data management activities required for big data—and you want to employ them in a variety of functions and units around your organization, I recommend the creation of this sort of job and title.
Targets for Big Data
Targets means that organizations need to select where they are going to apply big data and analytics within their businesses. At a high level, will the resource be applied to supply chain decisions, customer decisions, financial decisions, human resource decisions, or some other area? One cannot simply do everything with analytics at once, so targeting choices is a necessary process.
However, targeting has been more of a focus for conventional analytics than for big data in the short history of their coexistence. With big data projects, many organizations are just trying something to see if it will work; they are at the proof-of-concept stage. The projects are often picked because they are convenient or because the owners or stakeholders are willing to experiment. Rarely do organizations make a concerted attempt to determine what the most important or strategic project would be before beginning work on something.
Despite this lack of attention to targets thus far, it’s clear that an organization can’t deal with all big data at once, nor apply it all at once to all the areas of the business that might benefit from it. Targeting, then, is a necessary activity. Your management team needs to come up with answers to some of the following questions:
• Where do we have significant data resources that are unexploited?
• Which of our business processes is most in need of better decision making?
• Where would we benefit from much faster decision making?
• How might we create data-based products or services, and in which parts of our business would they be most relevant and useful?
• Is someone else in our industry likely to employ big data in a way that will disadvantage us? If so, how are they likely to use it?
Since the exploitation of big data can involve new products and services (in addition to internal decision support), there may well need to be more integration of big data initiative targeting with product development and strategy processes. If you’re developing a new product, can there be a big data adjunct to it—perhaps in the form of a service? If you’re thinking about disruptive innovations in your industry, how might big data contribute to them?
Analysts for Big Data
The only issue to address here is whether smart human analysts are any more important with big data than they were with traditional analytics. There is probably more emphasis on data scientists in big data than there was on quantitative analysts for analytics in the “old days.” But that’s not because they’re more important; they’re just harder to find, and they have a sexy new job title. The combination of technical and analytical skills required to be a successful data scientist makes them somewhat rare and difficult to recruit. Otherwise, you’ll need smart, capable people to do business analytics, regardless of what type of data you are employing. If you’re building your business around big data and analytics, like Google, GE, LinkedIn, and other household names, you’ll need hundreds of them.
Other Factors to Consider in Big Data Success
Is there a big data culture? Or put more precisely, is a big data culture different from an analytics-oriented culture? The differences are subtle, and a firm that wants to succeed with big data could do much worse than to adopt a culture that emphasizes analytical and fact-based decisions., Below are some of the attributes of a big data culture that I’ve observed.
Impatience with the status quo, and a sense of urgency: Among the data scientists and company leaders I interviewed, there was a strong belief that the big data market is a land grab, and to the early movers will go the spoils.
A strong focus on innovation and exploration: Big data firms are constantly innovating, exploring, and experimenting to learn more about their operations and their customers. Google was perhaps the first big data firm, and it sets the tone for culture. It encourages every employee to be innovative and gives engineers a percentage of their time to work on new products.
Belief in technology as a source of disruption: For many early adopters of big data, technological innovation is as important as data innovation. Google and Amazon push the frontiers of not only software, but also hardware and data center technologies.
A culture of commitment: The leaders in big data are willing to commit to bold, audacious goals. Google committed to making the self-driving car a reality—a project it considers to be based on big data. If your company hasn’t even discussed big data at the senior management or board level, you may want to address that situation.
Nonhierarchical and meritocratic organization: Big data early adopters believe that big ideas can come from anywhere and anyone in the organization. I’ve already discussed how LinkedIn’s Reid Hoffman empowered Jonathan Goldman to develop the People You May Know feature.
Embedding Big Data
How should you embed big data, and the analytics based on big data, into key operational and decision processes? I’m not talking about fully robotic (machine-only) processes—humans can still review and override the recommended actions from such automated systems— but it is important to employ automation to ensure that analytics and big data are employed in a timely and efficient fashion.
As big data and analyses based on it proliferate in society and organizations, we simply won’t have the time or the human labor to present those results to humans for decision-making. And we know more and more about the irrational decision processes that many humans employ; why wouldn’t we prefer more automated analytical decisions in many cases?
Analytics in the small data era were just beginning to become more automated when big data came along. Now we have no choice but to embed big data–based analyses into business processes.
For example, Heathrow Airport has been implementing a semiautomated system for managing flight operations—called Airport Collaborative Decision Making (A-CDM) for several years. Using a set of decision rules and process flows, this system automatically creates and dynamically coordinates all operations involved in flight turnaround, including exactly when a plane will land, which gate it should taxi to, how much baggage needs to be offloaded, when refueling will happen, the arrival time of the next flight crew, the time passengers will board, and when the plane should push back and take off. After two months, Heathrow had immediately improved the percentage of on-time departures from 60 to 85 percent -it would simply be impossible for humans to deal with all of this data without an automated process.
It’s clear, then, that the stand-alone era for big data and analytics is approaching the end of its useful life. Will the last human being to leave the big data building please turn out the lights?
Excerpted from Big Data at Work: Dispelling the Myths, Uncovering the Opportunities by Thomas H. Davenport. Copyright 2014 Harvard Business School Publishing. All rights reserved.
About the Author
Thomas H. Davenportis the President’s Distinguished Professor of Information Technology and Management at Babson College, a fellow of the MIT Center for Digital Business, cofounder and Director of Research at the International Institute for Analytics, and a Senior Advisor to Deloitte Analytics.