WHITE PAPER • SEPTEMBER 2013 Emergence of the Chief Data Officer: Leadership in the Age of Big Data

BY ADAM CHARLSON, EXECUTIVE & MANAGING DIRECTOR OF THE WEST COAST KRISTEN BARGE, PRINCIPAL

Copyright © 2013 DHR International, Inc. All Rights Reserved. Emergence of Chief Data Officer • Executive The rapid evolution of ‘big data’ has opened a new frontier where Summary organizations seek competitive advantages that did not exist before the digital era. However, to develop a successful big data program, companies need skilled data scientists and leadership that can oversee company-wide big data initiatives.

Just as the chief information officer (CIO) emerged in the 1980s to lead the mass adoption of IT systems, and soon became widespread, we predict that the chief data officer (CDO) will become a universal and highly valued member of executive teams. Companies in retail, manufacturing, agriculture, resource extraction, finance and professional service and tech, among others, will join the likes of Google, LinkedIn and Facebook, all of which currently employ a CDO.

The costs of data collection and storage, the increasingly scientific nature of data analysis and the high risks of big data programs demand not only time, resources and careful strategy, but an entirely new data-driven culture. Consequently, big data initiatives will need the visibility, advocacy and guidance that only an executive leader can provide.

The chief data officer will build a data-driven culture where accurate information is gathered ethically, shared widely and used methodically to grow profit. The CDO will also construct a big data ‘ecosystem’ to structure data that would otherwise exist in siloes.

The CDO and his or her team will combine the scientific methods of an R&D team with the business mindset of strategy consultants. The CDO will pinpoint promising research areas, propose shifts in corporate strategy, identify threats and oversee the creation of internal and external facing systems that rely on big data.

To gain a foothold in the big data frontier, and profit from it, organizations will need CDOs with a rare combination of specific technical knowledge and superb interpersonal skills. As companies continue to recognize the immense value and opportunities in big data, competition for data scientists with the right blend of hard and soft skills has increased, and will continue to increase.

The revolution of big data requires the evolution of executive teams that wish to profit from it. Throughout the business world, the emergence of chief data officers will herald this transformation.

Every day, businesses, governments and researchers collect billions of data points on consumers. This continuous and ever growing process fills countless servers with information localized by space, time and measurable behaviors. Indeed, the International Data Corporation (IDC) reports that the digital universe will grow 300-fold between 2005 and 2020 to a total of 40 trillion gigabytes. Yet, according to the IDC, only one percent of this data is currently being analyzed.

Copyright © 2013 DHR International, Inc. All Rights Reserved. Emergence of Chief Data Officer • 1 Therefore, organizations that are capable of gathering, structuring and analyzing data stand to gain a competitive edge in their respective marketplaces. The race for big data is on, but many organizations now wonder: Who can lead a big data initiative? Who is capable of converting billions of gigabytes into billions of dollars?

In the 1980s, as companies began to adopt personal computers, servers and digital technology en masse, a similar question arose. To leverage technology, organizations needed a leader who could manage the deployment of productive, cost-efficient IT systems. To solve this dilemma, forward thinking organizations began to hire chief information officers. By the 1990s, the CIO was ubiquitous on executive teams.

Today, we predict a congruent trend with chief data officers. As companies seek to identify, collect and leverage big data, they will hire leaders with the skill and experience to guide this process. Like the CIO, the emergent CDO will be found across industries. Companies in retail, manufacturing, agriculture, resource extraction, finance and professional service and tech, among others, will join the likes of Google, LinkedIn and Facebook, all of which currently employ a CDO.

Organizations on the edge of expanding their big data program or hiring a CDO should consider five key questions:

1. What are the goals and characteristics of a big data program? 2. What are the responsibilities and duties of a CDO? 3. What skills and background should a CDO possess? 4. Where will an organization find data scientists? 5. How does a CDO create value and drive revenue?

This paper will help executives, board members and investors conceptualize a big data program, understand the value of a CDO and see real situations where data science has produced value. Our research draws from open source publications, experience helping organizations find qualified data scientists and interviews with data scientists and data providers, including Dr. Partha Saha, Chief Data Architect for Data and Analytics in Online Services at Microsoft, and the CDO of a Fortune 500 retailer, who wishes to remain anonymous.

The Nature of Data The big data division of an organization should include highly skilled mathematicians, Science in Business computer scientists, academically trained scientists, hackers and software engineers who use the scientific method to create value for your organization. Unlike marketing, sales, operations, accounting or human resources, which deliver revenue growth in fairly established ways, a big data program creates value by utilizing information collected in every division of the organization. Compared to other teams, your data group’s project cycles could be significantly longer and somewhat riskier.

Copyright © 2013 DHR International, Inc. All Rights Reserved. Emergence of Chief Data Officer • 2 Data scientists will thrive in businesses that are “instrumented,” according to Dr. Saha. An instrumented company simply collects data on everything possible, and its data teams develops a system for categorizing, labeling and storing data in a way that makes it easily retrievable and relevant for answering any business question. A data science team will be acutely interested in this collection process because they need quality, accurate data to get good results.

Data scientists function like scientists in academic fields: they will ask a question, develop a hypothesis, collect data (often through an experiment), analyze the results and then report their findings. Their reports should include strategic insights that other departments will have to implement.

Data scientists do not ask, “How many people clicked this button?” That is simply one piece of raw, unstructured data. Instead, they might ask, “At what discount percentage is a white male, 18 to 25 year old shopper most likely to add an item presented at the initial checkout screen?” Then, the data scientists can ask which items to present, where to locate the offer button on the screen and ultimately create revenue where it did not exist before. In large businesses, convincing a small percentage of shoppers to add one item to a shopping cart could make a tremendous profit.

Thus, data scientists often show other departments the unintended negative consequences of their actions, or what they could be doing better. They develop algorithms and models that might be used directly in a recommendation engine or an inventory system. In a sense, they are a hybrid between R&D and strategy consultants that leverages a corporation’s full array of data instruments and test environments to create value. Like R&D and consulting, however, they also require a technological and financial investment.

As Dr. Saha explains, your data science team wins a competitive advantage for your organization in at least four critical ways:

1. They can observe current trends and patterns in business and get ahead of them before they repeat.

2. They can cost-efficiently interconnect with internal business and external public datasets to get more complete and therefore true .

3. They can help the business with data-driven innovations to engage in effective one-on-one customer management cheaply through electronic media.

4. They can quickly help the business take advantage of any innovations taking place in the Internet world around big-data including new marketing methods, optimized marketing campaigns, action-based social campaigns, brand awareness and more.

.

Copyright © 2013 DHR International, Inc. All Rights Reserved. Emergence of Chief Data Officer • 3 When an organization asks “Is this effective?” the data scientists can provide a data-driven answer and solution. They have, as our anonymous CDO claims, “The freedom to think across domains and outside systems.” The best data science team is both an antidote to repeated mistakes and an innovator of solutions and systems that harness big data.

Responsibilities and The chief data officer oversees big data collection, use and strategy throughout an Duties of the CDO organization. The CDO’s mission is to create a data-driven culture where accurate information is gathered ethically, shared widely and used to grow revenue and cut costs. As both Dr. Saha and our anonymous CDO argue, the CDO must ultimately create a big data ‘ecosystem’ that relies on existing data assets and acquires new assets in order to weave, connect and bring structure to data that would otherwise exist in siloes.

In an anti-data culture, employees form opinions and then find statistics that support pre- conceived conclusions. In a data-driven culture, employees ask questions, form hypotheses and then test their hypotheses against the data. A data-driven culture is comfortable with ambiguity and being wrong. A CDO’s responsibility is not to be right or to make other people look bad—the CDO’s job is to test the assumptions, procedures and beliefs that, if verified or challenged, would have the greatest effect on short-term profit and long-term health, and sometimes even the trade-off between the two. Your CDO must be intensely curious and unafraid to challenge his fellow executives. The most effective CDOs are brave, unconventional thinkers.

The CDO ideally oversees a team that handles all data analysis, from social media, marketing and advertising to pricing, customer service and operational processes. The CDO and his or her team is a fountainhead of intelligence and solutions that make every department more effective. The CDO will identify promising research areas and pathways to revenue, suggest shifts in corporate strategy, alert the executive board to threats and opportunities and oversee the back-end development of internal and external-facing systems that rely on big data.

The CDO also provides ethical oversight over data collection. He or she ensures that data is collected transparently and stored securely. The CDO and CIO will likely partner in this initiative, as collection, storage and legal compliance typically fall within the CIO’s purview.

Critically, the data scientist will also set the expectations of company-wide data initiatives. As Dr. Saha explains, “The data scientist should also raise awareness that ‘big data’ has a risk and reward, and some amount of failed ‘experiments’ are necessary to get things right. As such, the CDO is an investment in ‘innovation.’”

Copyright © 2013 DHR International, Inc. All Rights Reserved. Emergence of Chief Data Officer • 4 The Skills and The CDO must be capable of relentlessly identifying, collecting, structuring and analyzing Background of a data that is meaningful and actionable. CDO Dr. Saha says that “a CDO should be an innovator first, a scientist second and an active and fast learner either of industry or big-data practices. He or she is a chief scientist for the company.”

In a data-rich ecosystem, a CDO should possess the following technical traits: 1. Deep experience with big data solutions, including Hadoop or HBase.

2. Knowledge of data collection, classification, warehousing, architecture, integration, quality management and a very large (VLDBs).

3. Keen understanding of modeling techniques.

4. Familiarity with business and academic research methods.

5. Detailed knowledge of how data solutions are implemented in existing programming languages and design patterns.

6. Ability to train a data science team to efficiently solve business problems.

The CDO must also be business savvy. A CDO with scientific disposition but limited understanding of financial management, business processes, and compliance issues may lead a data project into legal trouble or grapple with research questions that do not increase the bottom line.

Without question, the most valuable quality in a CDO is superb soft skills. A CDO without the ability to communicate findings, win supporters and spur people to action will simply play a “cosmetic role,” according to Dr. Saha. Because a CDO will often have to question or challenge other executives, the CDO must make clear that his or her interest is the success of both colleagues and the company. The CDO must be able to enroll disparate departments in the sharing of data, which ambitious VPs and directors are often reluctant to do. A CDO without soft skills may cause friction in upper management and create political divides and mistrust between department heads.

Sources of CDOs We recommend that corporations consider candidates who have experience at companies with reputable data programs. Yahoo!, Twitter, LinkedIn, Netflix, Amazon, Google and Microsoft are all renowned for their big data programs. Companies might also promote VP or director-level employees in data management, business intelligence, data architecture and services.

Copyright © 2013 DHR International, Inc. All Rights Reserved. Emergence of Chief Data Officer • 5 While no universities offer programs in data science yet, DJ Patil, one of the co-coiners of the term “data scientist” and the data scientist in residence at Greylock Partners, suggests that organizations look for candidates “from the usual suspects” - Stanford, MIT, Berkeley, Harvard and Carnegie Mellon—but also a few proven wild cards like North Carolina State, UC Santa Cruz, the University of Maryland, the University of Washington and UT Austin. Patil suggests that some of the best data scientists have PhDs in esoteric fields like astrophysics, ecology and systems biology

With wide exposure to a specific industry and training in business skills, a new hire with strong scientific credentials can eventually develop into a candidate for the executive board.

The Value of a The value of CDO should be unmistakable. Whether they develop an algorithm that CDO: Case Studies predicts the lifetime value of customers (LTV), create a recommendation engine on your e- commerce site, forecast raw material demands or write an algorithm for fraud detection, the fruits of their curiosity should be obvious. The competition page on Kaggle, a platform for predictive modeling competitions, can provide an idea of the types of problems and challenges that organizations are currently posing to data scientists.

To illuminate ways a CDO could deliver value to a corporation, here we share three cases where data scientists demonstrated remarkable ingenuity and delivered revenue-driving insights.

Google Searches and Market Moves In April 2013, Tobias Preis, a business professor, and two physicists published a paper titled “Quantifying Trading Behavior in Financial Markets Using Google Trends.” The researchers found a remarkable relationship: as the volume of searches for words like “debt,” “portfolio” and “stocks” fell, the Dow Jones tended to rise. Indeed, a stock trader using what the authors term “a Google Trends strategy” based purely on search volumes of “debt” would have yielded a 326 percent return between 2004 and 2011. Unsurprisingly, the BBC reports that the three researchers were approached by executives in the financial industry.

Amazon: the Shopping Cart Recommender Greg Linden, formerly a data scientist at Amazon, wanted to make shopping recommendations based on items already in a shopper’s cart. As Linden writes in his blog, “we had an opportunity to personalize impulse buys.” Linden coded a prototype and modified a test site of the Amazon.com shopping cart page so that it would offer recommendations based on his algorithm. A senior VP of marketing said it would distract shoppers from checking out and forbade Linden from working on the project. Linden, however, wanted to measure the sales impact and won the support of executives. In tests, the shopping cart recommender was a tremendous success—in fact, Amazon realized that not having such a recommendation engine was costing the company money. Linden went against orders, but he tapped into his coding skills and soft skills to win support and prove that his recommendation engine would create value.

Copyright © 2013 DHR International, Inc. All Rights Reserved. Emergence of Chief Data Officer • 6 Gilt Groupe: One-on-One Communication

In “Big Data, Analytics, and the Future of Marketing and Sales” an e-book from the McKinsey & Company, Alexis Maybank, co-founder of Gilt Groupe, explains how the company used big data to reach $500 million in sales in just five years. The company mines extensive customer data to personalize all e-communication. A consumer on Gilt Group’s email list, for instance, will receive one of 3,000 different messages depending on his or her expressed interests. The messaging system is a joint project between company data scientists, who built the algorithm, and creative teams, which create imagery and language that will appeal to each segment of customers. The result is an intelligent, personal invitation to shop for goods that the customer is likely to find appealing.

Conclusions Globally, businesses are tuned into the immense value of big data and the rarity of data scientists who can wield it.

For large corporations, the stakes of the ‘data race’ are substantial. A report from McKinsey & Company, “Big data: The next frontier for innovation, competition, and productivity,” estimates that a retailer using big data can increase its operating margin by more than 60 percent—however, by 2018, the US alone “could face a shortage of 140,000 to 190,000 people with deep analytical skills as well as 1.5 million managers and analysts with the know-how to use the analysis of big data to make effective decisions.”

Organizations are now moving quickly to build data science teams and hire managers, directors and executives with the skills and experience to lead a big data program.

Given the high costs and risks of data science, the lengthy project cycles and necessity for some failures, we believe that organizations will increasingly demand the oversight, strategic direction and accountability that a chief data officer can uniquely provide. The role of big data is simply too big for the plate of a chief information officer or . While these executives will surely partner with the CDO, most CIOs and CTOs do not possess the knowledge and experience to guide a big data program, and most do not have the bandwidth to cultivate and champion a data-driven culture.

As universities begin to create data science programs and as companies with big data programs continue to train and develop data scientists, the pool of candidates will grow, but not nearly as fast as demand.

As this paper has illustrated, companies that can find a qualified CDO and build a data science program stand to become more competitive, innovative and nimble. Indeed, data science teams will be more than able and willing to measure their value to the company.

Copyright © 2013 DHR International, Inc. All Rights Reserved. Emergence of Chief Data Officer • 7 Established in 1989, DHR International is one of the largest retained executive search firms in the world, with more than

50 offices around the globe. We conduct search assignments at the board of director, C-level and functional vice president levels. DHR’s renowned consultants specialize in all industries and functions in order to provide unparalleled senior-level executive search, management assessment and succession planning services tailored to the unique qualities and specifications of our select client base.

DHR International

P 312.782.1581 • F 312.782.2096 www.dhrinternational.com

Copyright © 2013 DHR International, Inc. All Rights Reserved. Emergence of Chief Data Officer •