Support Learning and Insight

It’s more important than ever to understand the challenges facing financial executives. Support the Financial Education & Research Foundation today.

Strategy

Big Data May be the CFO's New Best Friend


by FEI Daily Staff

When it comes to information technology investment strategy, CFOs have more influence than any other executive.

© Federico Caputo/iStock/Thinkstock

Chief financial officers (CFO) are no longer just the financial stewards for their organizations. They're business strategists. And when it comes to information technology investment strategy, CFOs have more influence than any other executive, according to the 2013 Gartner FEI (Financial Executives International) CFO Technology Study.In today's business world, a comprehensive IT investment strategy must take into account the power of "big data."

Though it can seem nebulous and unstructured — and outside the traditional purview of finance leaders — big data is in fact one of the most important assets financial executives have when it comes to boosting productivity or profitability.

Big data is an aggregation of demographic information gathered from social sources like Twitter feeds, Facebook statuses, blogs and surveys, as well as automated tools like sensors. It can include government data, weather patterns and more complex variables. It gives insight into customer preferences, behaviors and purchasing patterns along with internal use of space and resources.

The analysis of this data presents an opportunity for CFOs to collaborate with chief information officers (CIOs) to improve business processes and outcomes. In fact, in a recent survey conducted by Oracle Corp. and Accenture PLC, 57 percent of finance executive respondents agreed that investments in big data and analytics will be key sources of competitive advantage moving forward. The problem is that big data is just that: big. Companies already deal with massive volumes of information. Consider, for example, Google, which handles 1.2 billion Internet searches every day. This level of data can quickly overwhelm finance leaders and their teams if they don't have the right IT infrastructure, processes and talent in place to store, process, evaluate and apply big data most effectively.

And "big" is only getting bigger. The McKinsey Global Institute estimates that data volume is growing by 40 percent each year and will increase by 44 times between 2009 and 2020.

Beyond the issue of volume is variety. Traditionally, enterprises mainly managed only structured data, which had a well-defined format that made it easy to categorize, search and analyze. Big data, however, is "messy." The fact that it comes from everywhere — from websites to sensors embedded in electrical networks — makes it very unstructured, yet also very valuable if mined for patterns and meaningful themes.

Finally, financial executives need to contend with the rapid speed at which big data operates. As the cost of mobile connectivity continues to drop and sensor networks become ubiquitous, businesses can collect data faster and faster, accumulating more and more of it by the second. The proverbial stacks of data can pile up much faster than any team of professionals can sift through them, especially without the right analytical tools in place.

As data grows in volume, variety and velocity, it transcends the capabilities of the traditional analytics, upon which finance executives and their colleagues may have relied. It becomes too rich (each data element is larger or more varied because it combines information from structured and unstructured sources), more granular (time intervals in data sets decrease from months to days or even hours or minutes) and it requires much faster processing. Integrating Big Data, Strategy and Financial Management There are steps finance executives can take to help simplify the process of integrating big data into the larger strategic and financial management picture. Using the following five-step action plan can help any CFO create the right environment to optimize big data experimentation and return on investment.

Step 1) Invest in Analytical Skills to Prepare for New Data. As noted above, big data brings with it more volume as well as more sources of information. Many of these sources live outside the enterprise and may therefore be less familiar to financial leaders and harder to organize. They have strong business value, however, when combined with financial metrics. CFOs face the challenge of learning what questions to ask when it comes to creating the right data analytics and applying them to business practices. McKinsey Global Institute also reported there will be a shortage of talent necessary for organizations to take advantage of big data. By as early as 2018, the United States alone could face a shortage of 140,000 to 190,000 people with deep analytical skills. Financial executives can, however, enable their teams to put big data to pro­­ductive use by investing in analytics education for top talent. The New York Times recommends programs at universities including Columbia, The University of San Francisco, New York University, Stanford, Northwestern, George Mason, Syracuse, University of California at Irvine and Indiana University, with more and more certificate courses "heavy on data" cropping up constantly.

Step 2) Conduct rapid experiments: 5-5-5. Big data is about reducing the time, cost and effort to form and test hypotheses. Unlike past processes, which involved millions of dollars of upfront investment in data warehouses and collection models, big data puts the data itself first. The collection of the information requires less formal structure because the information comes from such diverse, often unstructured places as weblogs and social networks. The combination of any of these sources makes for difficult, even impossible modeling, at least in the initial phases of analysis. More importantly, the value of a new data combination is often uncertain, making even a three-month data mart project prohibitive, never mind a 12-to-18 month data warehouse initiative. So how should financial executives approach big data and work with this raw information? Andrew McAfee, author, big data expert and principal research scientist at the Center for Digital Business at MIT Sloan School of Management, suggests a 5-5-5 approach for experimenting with big data: take five experts, give them five weeks and look to combine five data sources to discover new patterns or ideas about a business.

Step 3) Build Strategy for Valuing Data. All other assets on the balance sheet have a plan. Why not data? Because until now, it's been too difficult to properly value data. Valuation requires insight into the past contributions of data to the enterprise, and big data technologies now provide this perspective. With big data, organizations can track the influence certain information has on business decisions and determine whether that information led indirectly or directly to financial gain. And there is another challenge: as with equities or research and development (R&D) investments, past performance is no guarantee of future performance. Companies will have to design new ways of pricing the option value of information. To start this process, financial executives can track the value created by the 5-5-5 approach and experiment with algorithms to predict the observed performance.

Step 4) Data-First and Model-First Worlds Collide. When incorporating big data, the first challenge financial executives will face is how to efficiently collect and store large amounts of information. Initially, the grouping, aggregation, classification and filtering of millions of rows of data can be much easier if plugging the information into a set model isn't also a concern. Over time and with increased usage, the value of particular big data combinations in the data-first environment should become evident. At that point it would make sense to convert these tried and true combinations into data models. The data-first and model-first worlds are complementary, each producing information for use in the other. Finance executives can challenge the IT department to systematically convert big data discoveries into new operational best practices and turn data from operations into raw material for new discoveries.

Step 5) Augment Analytic Capabilities. Many of the enabling technologies for big data started in the open source community, leading to the creation of several big data working groups and boutique firms. To accelerate learning within their organizations, financial leaders can tap into the knowledge of some of the 80,000 data scientists available at online analytical community sites like Kaggle.com. Or, financial executives can hire specialist consulting firms. CFOs surveyed for the 2013 Gartner FEI study emphasized the need for focused investment in business intelligence and analytics. By following the suggestions above, finance executives can begin to put this investment into action through collaboration with CIOs and other members of the management team. The more insight and understanding CFOs can gain about their business through big data, the more they can help their organizations meet vital business objectives. With a clear and actionable view into big data, CFOs can help increase efficiency, improve collaboration and alignment between finance and the business, improve organizational agility and foster innovation. So, pay attention CFOs: big data may just be your new best friend.

Rich Clayton is vice president, business analytics product group, at Oracle Corp. in San Francisco, Calif.