How can organizations create business value from this glut of data? Success requires managerial talent who have the powerful combination of business savvy, leadership skills and analytical expertise. They need to recognize both opportunities and problems and have the knowhow to bring together technical experts with functional managers to drive results.
Against this backdrop, the Innovator convened three leading big data experts from the Graduate School of Management to share their insights and perspectives on how executives, companies and universities are pushing the envelope at the intersection of analytical thinking and information technology:
UC Davis alumnus Andrew Barkett MBA 09, formerly of Facebook, is chief technology officer for both the Republican National Committee and Data Trust, a firm founded in 2011 by former party officials and contracted to improve the party’s vast voter file and foster a culture of data and learning. Barkett was featured in a New York Times story about his lead role in developing a platform that will allow Republican candidates and state parties to easily swap and share data on Republican voters and create other analytic innovations.
Professor Naik develops new models and methods to improve the practice of marketing, using big data sets from industry in his research on marketing and advertising strategy, integrated marketing communications and dynamic market response models. His groundbreaking research gives managers powerful tools to determine the most strategic media mix for ad campaigns and measure the effectiveness of integrated marketing communication. He has done extensive consulting for Fortune 500 companies, including big data analysis for sales and brand management projects.
Professor Bhargava is an expert in technology management and the information technology industry. He has studied the use of IT in clinical health care, and has worked on data-driven and analytical decision making in organizations. His recent work examines elements of competitive strategy for technology products, including pricing strategies, revenue models, product design and product mix strategies, technological compatibility and interoperability, bundling, and supply chain considerations. He has studied or commented extensively on products and practices of Google, Microsoft, Apple, Verizon, AT&T, Amazon and IBM. Bhargava holds the Jerome J. and Elsie Suran Chair in Technology Management at the School.
Saigal teaches MBA courses in Decision Making and Management Science, and Supply Chain Management. He has more than two decades of experience in business process efficiency consulting in Silicon Valley and India. His clients have included big data analytics powerhouses such as Netflix and Walmart, and traditional manufacturing leaders such as Hallmark, Intel and Nestlé. He has served as the executive director of the St. Stephen’s Institute for Management Excellence in New Delhi, India, which he helped establish and where he teaches global best practices in information technology, decision-making, and operations management to Indian executives.
Naik, Bhargava and Saigal got together in June at the Graduate School of Management.
The dialogue that follows is an edited excerpt from their conversation.
Naik – Sanjay, you have been working with data-heavy projects in your consulting for the last two decades. Has the emergence of big data changed how companies look at decision making?
Saigal – Absolutely. The first thing to remember is that there are many, many different kinds of data now available to decision makers than ever before. In addition to the ERP, which is Enterprise Resource Planning data that companies have traditionally had for the past few decades, there is all of the social media data that’s out there, there is all the non-numeric data. And they’re having to integrate all of these things into their decisions. The pressure to make decisions based on leveraging this data is enormous.
Naik – Does that mean business leaders have to be more comfortable with data analytics as well as entry-level and middle managers who work with data?
Saigal – That’s exactly right. Bob McDonald, the CEO of Proctor and Gamble, has said that analytics need to move into the center of business operations. When you read interviews with him you hear terms like ‘Bayesian analysis,’ like ‘decision trees.’ These have become an essential part of the toolkit of a very high-level leader. Not just the CIO, but also the COO and CEO.
Naik – Sanjay, data now arrive at regular frequencies and shorter frequencies than in the past. Managers and companies are going to make decisions more often based on that data, and keep changing the decision as they go along. Does that open up the possibility that they would be making bad decisions just because new data came in, which is different from something else, and they shouldn’t have changed it in the first place?
Saigal – Absolutely. In fact it’s remarkable what is happening in business decision making. The cultural shift that is happening now is towards evidence-based management. So you look at data, you construct hypotheses, you make some changes, and you see how it goes. I recently had a COO of a software company tell me that his new VP of marketing was working out so well because he does A/B testing for any major change that they make to their website.
So, in this context, failure is a given. Failure is normal. You fail early, you fail often, and you learn from each failure. And this is a big contrary shift in the way management gets done out there precisely because of the availability of big data.
Naik – So senior managers should have a greater latitude of acceptance to having such failures?
Saigal – That’s exactly right. There is no longer that kind of a pressure, even at the senior level, to be in errant; to know the answer the first time. You try it out, and very often—as in the Procter and Gamble case—they have a “decision cockpit” where you can work with raw data, with live data. They can try it out immediately, and see if something works or it doesn’t work.
Saigal – Hemant, you’ve been looking at the value of information to businesses—how they collect it, how they use it, how they manage it. You’ve been doing this for well over a decade. How does big data change this picture?
Bhargava – Big data has very interesting characteristics and I’m going to talk about four things: size, format, age and discovery of data.
So first let’s look at size. Big data is obviously big, and lots of data. That means not simply just millions of observations, but that we have today, hundreds and thousands of information attributes about each object being observed. That creates interesting challenges to statistical inference because with all these attributes you can easily infer things that do not quite hold up in the real world, or are not necessarily interesting and novel.
The second aspect is the format of the data that we work with. We went from a world where we primarily work with structure and numeric data, to nonstructured, non-numeric textual or multimedia data. That comprises text, images, videos and animations, and so on. For example, a police department may discover an increase in crime activity by monitoring images from a number of street cameras.
Or a healthcare system might discover trends in new diseases by monitoring search terms on Google. These are brand new opportunities for employing data and making new discoveries relevant to the operation of business.
The third is the fact that the age of the data we’re operating on can be extremely young. In the past, firms took many months or years to run a full, decision analysis project from collecting data to cleaning it up and loading it in data bases, building models, running inference, learning new insights and then finally taking actions based on these insights. By the time you take these actions, the data that at you discovered the insights from are quite stale. And today we’ve collapsed that cycle time of analysis down to minutes, days and on many occasions even real time.
Fourth is the aspect of discovering the data that you’re going to work with. Today we have made a lot of progress in not just collecting data and putting them up for access online, in the cloud, but even putting up descriptions of the data—the meta information—in formats that machines can automatically find and understand what the data are about. So that an analyst or, or a data analytic system, can actually go out and discover data in real time. What we really have is a virtual supply chain of data coming in from many, many different sources and being put together just in time, on demand.
Collectively—the size, format, age and discovery—give us a world of data analytics that simply wasn’t there even 10 years ago.
Saigal – Prasad, how has the emergence of big data promised to change the teaching of management?
Naik – Management education is going to change dramatically as we go forward. Current management science relies on social sciences as a foundation of paradigms. Economics, sociology, psychology—they provide the basis.
There have been many refinements on this paradigm over the last 50 years. But, big data is going to change it dramatically into a new direction of STEM [science, technology, engineering, and mathematics]. Managers will need to think like scientists to do experimentation and try new hypotheses, try new actions and see what happens—and learn from that.
They will have to be competent in technology, where they can capture and manage data from multiple sources. They have to think like engineers, and have an engineering mindset to solve problems and build systems. And finally, they have to think like mathematicians, to instill logic in their reasoning of why some actions lead to certain outcomes.
And all these STEM-related sciences are going to impact how managers think. And that’s going to train them to be better managers and better leaders.
Saigal – Hemant, you’ve been working on big data-related initiatives at UC Davis and inside the Graduate School of Management. Can you share details about this work?
Bhargava – We are very excited at UC Davis about all issues around big data. This year the university created a Data Sciences Initiative, which is an umbrella to bring in various disciplines that are working with lots of data and analytics. And this was set up with very strong support from campus leadership, so it’s a very exciting development.
At the Graduate School of Management, we are working on creating a new, one-year graduate degree program in business analytics that is focused on the applications of big data in business. We want to design this program around four strong pillars: data management, analytical methods, knowledge of business, and organizational effectiveness. We want graduates to excel at formulating and executing strategic business analytics projects.
We also expect this program to be highly collaborative with industry, so that students can work on actual company projects and learn how to apply the new techniques around data to solve real problems in business. Right now we are developing the curriculum, setting up the program, establishing corporate partnerships and understanding what the projects would look like.
We are on the ground floor with regard to designing this program—and it will be unlike any other because we are drawing from expertise across the UC Davis campus. We would love to hear from our alumni and other stakeholders who can help us put together different pieces of this program.