Most companies focus on the past and present simply because that's all their technology allows them to do. BI tools generally provide little or no insight into the future. And in the quest to supercharge results and embrace business performance management (BPM), that's simply not good enough.
That's why many organizations are turning to predictive analytics. Predictive analytics capabilities, which are being incorporated into software products that generally fall into the categories of BI or BPM, allow an organization to run highly detailed simulations and develop appropriate strategies to execute throughout operational areas as well as within the finance function.
Here's how predictive analytics works. An application scans data sets and examines patterns that relate to successes and failures. Then, if sales dip below a prescribed threshold or the company exceeds its budget, the software not only pinpoints the specific problem, it sifts through the data to find contributing factors. Once the system identifies key issues, it analyzes an ongoing stream of operational and transactional data to provide alerts and insights.
Using a graphical dashboard, a predictive analytics application displays the probable impact of various scenarios. Not only does the software track how KPIs mesh with other metrics, they show how all these factors drive business processes. As a result, it's possible to notify managers about emerging issues and problems early on so that they can make quick adjustments. No less important: The system ensures that all managers abide by the same set of standards, and it can suggest a course of action based on previous circumstances and results.
Not surprisingly, numerous software vendors have streamed into the predictive analytics market. Several companies that have traditionally offered BI and BPM products, including Cognos Inc., Hyperion Solutions Corp. and SAS, have expanded into predictive analytics. Others, such as SPSS, continue to expand capabilities. Meanwhile, OutlookSoft Corp. has introduced a unified predictive performance management product that offers built-in strategic planning, budgeting, forecasting, statutory consolidation, reporting and KPI analysis. It is designed to deliver actionable information to the desktop and provide a complete view of the business.
What makes the combination of predictive analytics and BPM so powerful is that it provides an organization with a comprehensive and forward-looking view of key business factors. A company is able to "measure its execution against established metrics and performance goals and take immediate action," says Colin Teubner, an analyst at Cambridge, Mass.-based IT consulting firm Forrester Research.
While organizations have always relied on projections to run a business, predictive analytics takes the concept to a higher level. "Predictive analytics transforms BPM from a historical tool into a solution for planning the future," says Craig Schiff, president and CEO of BPM Partners in Stamford, Conn.
Indeed, a problem with traditional BI and BPM is that they track performance for prior periods and display organizational performance in the rearview mirror. Moreover, it's not unusual for all the data residing in these systems to prove overwhelming and bog down decision-making. On the other hand, when an organization taps into historical data and plugs it into predictive analytics tools, it's possible to evolve from a reactive to proactive mind-set.
Predictive analytics helps executives understand the probability of a particular scenario unfolding -- whether it's meeting revenue numbers or achieving customer-service objectives. At that point, the enterprise can plan for various outcomes. Simply put, it transforms linear thinking into three-dimensional planning. "Today, companies realize that even the best strategy will not stand up over a five-year span," says John McMahan, senior business adviser for The Hackett Group, an Atlanta-based consulting firm. "They recognize that strategy is a reaction to the environment, to competitors and to changes in commodity prices."
The intersection of predictive analytics and BPM offers enormous potential for improving operations. Computer manufacturers such as Dell use predictive analytics to adjust product offerings and pricing in real time. Wal-Mart, Target and other retailers increasingly use systems to gauge consumption patterns and predict which products are going to sell at which store locations during which periods. Banks use predictive analytics to detect potential fraud and identify changes in customer behavior that could lead to transferring funds or closing an account. And airlines use these systems to gauge supply and demand, adjusting pricing dynamically.
Oil companies rely on predictive analytics to boost the odds of finding oil. "The oil industry very early on used predictive analytics for high-value decision-making. Where do we drill? If I can improve the decision of where I drill by just a small percentage, I can derive a great deal more value and revenue stream," says Thomas Jenks, practice director at Parson Consulting in Chicago.
Such capabilities are also filtering into finance. A growing number of organizations are turning to predictive analytics to improve the accuracy of revenue forecasts, manage budgeting and understand credit risks and the value of various capital investments more effectively. "Ideally, predictive analytics pervades every area of an enterprise," says Schiff. "It's something that should permeate performance management."
Yet the challenges of putting predictive analytics to work -- and reaping rewards in the BPM arena -- are significant. For one thing, organizations must define which factors to plug into a predictive analytics system. As always, too much data or the wrong data yields sub-par results. For another, many software vendors are only beginning to wrap their arms around predictive analytics and offer products designed for enterprise use. Finally, it's essential to build workflow and processes for managing the resulting data -- and improving the decision-making process.
Although predictive analytics -- particularly as it relates to BPM -- is in its infancy in a small number of companies, the trickle is likely to change into a torrent in the months and years ahead. That's because the combination of predictive analytics and BPM generates financial and practical gains that otherwise would lurk under the corporate radar. "Predictive analytics doesn't provide clear-cut answers; it defines possibilities," says Schiff. "It's up to people to analyze various scenarios and possibilities and choose the right business strategy. Nevertheless, this numerical approach is gaining favor."
Putting predictive analytics to work as part of a BPM initiative requires both technology and well-defined business processes. According to McMahan, the most successful organizations rely on 10 to 15 key indicators to drive decision-making. Using more than this number makes it difficult to sift through the data and obtain useful and actionable results. Too few KPIs and it's impossible to achieve meaningful results.
One company effectively navigating the intersection of predictive analytics and BPM is Brother International Corp. in Bridgewater, N.J. The U.S. division of Brother Industries Ltd., a $4 billion (2005 sales) manufacturer of printers, fax machines, sewing machines and other devices, began using predictive analytics in 2003 and has increasingly tied it into business performance management. Using SunGard AvantGard's Get Paid application, the company is able to compare outstanding credit and collections with business metrics and drill down through the data to examine individual customers.
That alone has helped boost financial performance, says Susan Delloiacono, director of credit for Brother International. "We're able to look at data by collection region and react far more quickly. We are able to manage accounts more effectively and make decisions about extending credit on a more informed basis." But Brother hasn't stopped there. The company is now using the data to analyze cash flow and revenues and adjust decision-making accordingly.
The result? The analytics application has improved decision-making related to borrowing and improved the accuracy of budgeting and revenue forecasts. In addition, the system has reduced the time it takes to open a new account from more than a week to 3 days. Delloiacono says that these systems have also helped the company cut costs by reducing the number of employees in the credit department by 25 percent. "We achieved ROI on the project within 90 days," she says.
The engine for predictive analytics is complex statistical analysis. Algorithms built into the software allow an enterprise to use various techniques to simulate real-world behavior and business conditions. Monte Carlo simulations, for instance, randomly generate values for uncertain events. For a food producer, this might include weather conditions, labor problems and crop failures. Other techniques, such as regression analysis methods (which measure the relationship between related variables) and R-Squared models (used to develop capital-asset pricing models) also play a valuable role in uncovering patterns.
"When an organization begins applying those kinds of statistically valid techniques, predictive analytics moves behind marketing and hype and enters the world of real mathematical science," Jenks says. At that point, it's possible to compare scenarios, view probabilities and engage in effective contingency planning. "The advantage to effective predictive analytics is that it takes a lot of the guesswork and subjectivity out the process."
A May 2006 study conducted by OutlookSoft and BPM Magazine found that 87 percent of respondents believe predictive analytics is important to the budgeting and planning process. Yet only 17 percent say that they are using the technology to drive improvements in these processes and tap into effective business performance management. "Most organizations recognize the power of the technology, but few have merged predictive analytics and BPM in any significant way," McMahan explains.
For finance, the opportunities are enormous. An initiative that combines predictive analytics and BPM can help drive financial best practices throughout the entire enterprise -- from sales to support to financial planning. This, in turn, can lead to streamlined business processes and greater shareholder value. It can also build a more agile and flexible organization that's able to adapt to change rapidly and dynamically.
That's a winning proposition -- and one that's likely to catch the eye of business decision-makers in the months and years ahead. Concludes McMahan: "The use of predictive analytics and BPM is the next frontier for many organizations. It ultimately brings automation and consistency to very complex planning processes…and it allows decisions to take place in real time."
5 Ways to Effectively Tie Predictive Analytics to BPMAn effective strategy for tying together BPM and predictive analytics centers on the following core issues. An enterprise must: Identify key metrics. Finance and other departments must determine which metrics and key performance indicators to analyze. The number usually falls between 5 and 15 and encompasses diverse metrics ranging from days sales outstanding to scientific data used for oil exploration. Build a BPM model. Success depends on developing a sound business model that incorporates data, rules, formulas, workflow and other factors. It's also necessary to build templates so that line managers and others can view relevant data and act on it as needed. Integrate applications. Analytics tools and BPM software must integrate with core systems, such as enterprise resource planning (ERP), material resource planning (MRP) and customer relationship management (CRM). Data must flow into the analytics application in order to produce valid results. This might translate into investments in middleware and programming. Train executives and managers. Without training, even the best predictive analytics system is likely to go astray. Decision-makers must understand how to view dashboards and act on data in a timely and effective manner. Monitor data and act on it. Finance can play a key role in understanding data sets -- particularly financial information related to BPM -- and putting analytics to use. It can also help other departments make data actionable. |