In 1973, the British economist E.F. Schumacher published a popular book whose title became a rallying cry for those who thought the overuse of technology was destroying the earth’s resources: Small is Beautiful. With much of corporate America on the verge of plunging into projects involving “big data,” finance chiefs are being advised to take Schumacher’s dictum to heart.
Urged on by Silicon Valley entrepreneurs, companies in industries ranging from health care to breweries are indeed looking into ways to turn their caches of information into money-saving or money-making assets. At a recent Bloomberg CFO conference, for example, Todd Papaioannou, chief technology officer for Splunk, a San Francisco-based, big-data-oriented software firm, warned finance chiefs to “compete on your data, or you’re going to die.”
Faced with such challenges, growing numbers of companies are looking to get on the bandwagon. Nearly half of 751 professionals responding to a recent IDG Enterprise survey who buy products and services for their companies are forecasting big data initiatives at them. A third say their companies are planning one within 24 months.
Still, only 12 percent of the companies represented have already implemented a big data project, and 39 percent have no plans for one. The rest—49 percent—range from the pilot-testing stage to the planning-but-no-timetable stage.
Those figures suggest that many companies are still at sea regarding the parameters of their projects. In such an environment, it’s crucial that companies keep their goals – and the scope of their projects – beautifully small.
Beyond the Sea
For example, one key point that CFOs should keep in mind when deciding whether to approve spending for a proposed Big Data project is to make sure the company isn’t trying to “model the ocean,” says Phani Nagarjuna, chief executive and founder of Nuevora, a big-data analytics and apps firm.
CFOs should avoid proposals like “we have all this data, and we want to analyze the flow of that data and see what is it that we learn,” Cautioning finance chiefs to keep their companies from “just getting into an analysis exercise for the sake of analysis,” Nagarjuna advises that “if they have to invest a single dollar into a big-data initiative, they have to have a goal.”
That goal may be, for example to understand what future cash flows will come from current and future customers or to understand the predictability of those future cash flows. Whatever the goal is, “it needs to be tied to a business outcome,” he said. “What is the business outcome that you are trying to optimize?”
CFOs should put that question to the person requesting a big data project, whether it’s the chief development, investment, information, marketing, sales or merchandising officer. The CFO should also ask: How much of a difference can the project make, and what does it mean for the top- and bottom-line growth of the company?
Part of the reason for the ill-defined nature of some big- data projects may be that the term means different things to different people. To be sure, it generally refers to data collected in large volumes to be analyzed via new technologies to extract economic or societal value. But the information can stem from many kinds of sources, including transactions, social and other online media, smart phones, unstructured documents and machines.
At companies with 1,000 or more employees, emails are the largest big data source, followed by customer databases, online portals, Word documents and transactional data, according to IDG.
Besides setting precise goals and identifying the prime sources of data for the project, management needs to “carefully plan out what you want to accomplish with [a big data] initiative,” says Chris Scott, a regional partner in charge of technology issues at Tatum Consulting.
Next, senior executives need to analyze what types of management personnel and infrastructure may be needed. And once the parameters and systems are in place, the data needs to be clearly communicated. “You can put a lot of data into a Big Data engine, but you need to do something productive with it, and you need to find a good way to present that to people,” Scott said, noting that visualization tools can be a help in presenting findings to users who aren’t big-data scientists.
Another consideration is being able to repeat the data-gathering and analysis processes, Nagarjuna said. The CFO should make sure that as different business units of the company embark on their own individual big-data projects, they’re linking the learning gained from the company’s first big-data project to the next one, and so on.
That drives down the cost of subsequent projects, eliminating the need to go back to the drawing board and ensuring that best practices adopted along the way will be repeated.
CFOs also need to consider whether their companies should use Big Data service providers or manage projects in-house, Scott said. A company that manages its project in-house has the option of using cloud- hosting services, he said. If the company uses an outside service provider to collect data, the CFO has to consider the potential costs of getting the data into the company’s own big-data engine.
While analytics providers might be helpful in providing useful data correlations, they may have little use for companies that want to dig deeper into the data. For its own big-data project, Merchant Cash and Capital of New York, an alternative-financing company for small merchants, hasn’t contracted with providers that analyze customer data, says Jeffrey Beckwith, the company’s CFO.
That’s because providers usually won’t give up raw data along with their answers, Beckwith said. “We want to maintain a database with all of the raw data so that we can do historical analysis,” he said.
“Instead of using some third party that’s providing us answers to certain questions using big data, we want to be able to take that data ourselves, store it, answer our own questions, and then if we need to somehow massage that data in a different manner, we then have the ability to do that,” the finance chief added.
Even for companies that are not yet using Big Data, CFOs should be evaluating their industry for when other companies might begin using it, Scott said.
One way to tell is to ask the question: Is there data that exists that you would like to use, but that you don’t have access to? In the area of professional services, for instance, does data exist that could answer the question: Why did someone purchase your service, and why didn’t someone else?
CFOs of companies mulling an entry into Big Data would do well to anticipate some of the issues that a project could raise. The biggest challenges faced or expected at companies with 1,000 or more employees, are the limited availability of skilled employees to analyze data and manage Big Data, according to the IDG survey. Other worries include limited budgets, security and legacy issues. Considering all the companies mulling projects, however, it’s unlikely that those barriers will stem the tide of big data.
Keith Button is a freelance writer based in Valley Cottage, New York.