Occasionally someone tries to funnel some of the hype and hope surrounding big data analytics into the drier realms of bean counters, budgeters and forecasters. But the information that the typical CFO handles on most days just doesn't seem "big" enough to merit big data. Little data is more like it. Even in the face of ready anecdotes about 10,000-cell Microsoft Excel spreadsheets, skepticism seems in order.
Yet recent chats with a prominent service provider and an industry analyst suggest that for the finance office, the truly big data will be all the information that companies collect internally during their normal business processes. By turning the lens of analytics toward workflows spiffed up with time stamps and airtight integration, CFOs can spot the places where resources leak out and opportunities for new revenue go unnoticed.
Eric Simonson, managing partner for research at Dallas-based Everest Group, a consulting firm that specializes in business services, agrees that analytics tools in CFO offices are typically used for more basic purposes than the hype suggests. But he sees "significant movement in this direction" and believes it is only a matter of time until most sophisticated organizations apply them to big data.
One reason for the lag, according to Simonson, is that for finance, the big data analytics opportunity is less about traditional financial and ERP systems -- which suffer from numerous limitations -- and more about processes where operations and finance intersect and generate lots of data, or can be appended with external data to generate insights.
Inside big data
With rare exceptions, until now the closest that big data has come to the CFO's office has been the well-publicized examples involving the enormous consumer databases of credit card vendors such as American Express, who use them to identify -- and sell data on -- buying patterns and credit worthiness, among other things. The typical finance department encounters nothing that approaches the size of those terabyte-scale databases when closing the books, preparing budgets and forecasting revenues. Simonson nonetheless sees similar fraud-detection and "risk/credit" applications eventually picking up among early adopters in corporate finance departments.
"The promise of big data and analytics for CFOs is first being seen in scrubbing financial transactions which are already cleanly captured and benefit from more advanced analysis to find oddities and patterns," Simonson wrote in an email interview. "For example, although individual payment transactions may each appear appropriate, advanced analytics can help identify areas that may represent fraud, which can then be investigated further. Similarly, credit risk can be better predicted by looking at large amounts of data, and more efficient analysis tools can help audit more transactions and even use unstructured data to identify potential outliers."
There's good reason to believe the know-how will come mostly from outside. That's certainly the contention of Tom Dobis, the global lead for the business process outsourcing (BPO) practice at Hewlett-Packard Co. Dobis, based in Geneva, Switzerland, sells BPO services for finance and accounting (F&A).
Like many other advocates for financial analytics, Dobis says you don't have to bring analytics expertise in-house, with all its hiring and training pressures. That's what outsourcing services like HP's are for. An outside firm that also has a strong history as a system integrator can provide the integration and workflow that ERP systems often lack, especially on the financial side, he said.
But isn't this mostly a luxury of well-heeled buyers? While it is true that the few early adopters have been large companies with more money to spend in-house, outsourcing makes such projects possible for smaller companies, Dobis claims. "A lot of midmarkets don't ask for it, and they could."
HP's business process management, workflow and document management engines, especially its AutoFlow tool, are the key, and they have been the focus of significant investment by the company, Dobis said. With the end-to-end, time-stamped workflows that these engines enable, companies can spot the root causes of operational glitches that cut into the bottom line -- for example, sales managers who are taking too long to approve deals, or invoices that stagnate in accounts receivable queues.
One workflow that HP has designed for customers is a dispute-resolution system that can help identify valid claims and make sure that customer service resolves them quickly so the most valuable relationships remain in good standing. Dobis said some people use payment disputes as a money-saving ploy, and a good analytics-enabled workflow can help spot such abusers, shutting them off before resources are wasted on them. Collections are another early application, and price-analysis tools that HP uses internally are also available to customers of its services wing.
More on financial analytics
Read a definition of financial analytics
Understand the analytical potential of XBRL
Learn about an Oracle financial BI application
Dobis swears this is all real, citing completed or pending projects at four customers who are using analytics for fraud detection, cash forecasting for working capital management, or payment recovery. Sixty more companies are keenly interested. And almost all of the 15 to 20 CFOs of major companies that Dobis met with recently named analytics as their immediate priority. HP competitors like KPMG and Deloitte are "all on to the analytics stuff," too, he said.
While these early successes hint at a promising future, the old hurdles remain, according to Simonson. "The ability of finance organizations to leverage analytics varies significantly based upon their investments to date in building large data sets and developing business intelligence tools," he said.
David Essex is executive editor of TechTarget's business application websites. You can reach him at firstname.lastname@example.org and on Twitter @dessexTT.
This was first published in August 2013