Yogi Berra got it wrong. For most companies, the future is what it used to be: a perpetually blurry hodge-podge of vague expectations, missed opportunities, and unpleasant surprises. Forecasting remains a blind alley.
Software companies sense an opportunity and are attacking the forecasting problem from a number of vantage points. Three-year-old UpShot Corp. began as a sales-force automation firm, but in June it added a significant new dimension to its Web-based software: a module that provides real-time forecasting. The company claims that for most businesses, the sales pipeline provides the clearest view of the future, but that a reliance on spreadsheets and other outmoded tools guarantees that what little data exists is outdated and inaccurate. The firm's InVision Forecasting module captures the progress and projections of the sales force and allows managers to drill through the information in real time. The software will even compensate for an overly optimistic salesperson: if promises have consistently surpassed results by 50 percent, the software allows managers to cut projections in half. One limitation of the product, of course, is that it does not provide visibility beyond the length of a sales cycle. But as Keith Raffel, the company's chairman and founder, says, "That's a significant amount of time for most companies, and can give a CFO greater predictability."
Later this year, UpShot will modify the product so that it can blend information from the sales force with customer information that resides in other systems. The ability to integrate various sources of information has been a selling point for a number of vendors in the budgeting and planning space (Hyperion, Comshare, and others) for some time. Delbert Krause, a director of product marketing at Cognos Corp., says that better forecasts are a grass-roots effort: give employees access to a Web-based system that allows them to enter different types of data, and they will be able to combine and analyze that data in a way that produces better forecasts.
Alphablox Corp. provides tools that let companies build such systems. Using Alphablox products, Deutsche Bank built a system that reaches into its data warehouse and CRM (customer relationship management) databases and facilitates "SWOT" analysis (the identification of strengths, weaknesses, opportunities, and threats). One goal, says Edwin Ball, Deutsche Bank's global coordinator of MIS, is to predict which types of sales efforts will prove most lucrative. All companies want to up-sell and cross-sell current customers, but most take a shotgun approach. Deutsche Bank uses its system to analyze the revenue potential of each client and model different sales approaches. "Many companies forecast sales by taking this year's revenue figure and adding an arbitrary percentage," says Ball. "We're moving toward a system in which we augment internal data with market- share rankings, surveys, and other information."
Alphablox CEO Polly Sumner says that with the economy in a slump, forecasting becomes more important. "CFOs can't just cut costs," she says. "To manage through a downturn, they need to be able to see and respond to rising demand."
Gartner research director Lora Cecere argues that the 80/20 rule applies to forecasting: 80 percent of the effort should focus on cross- functional communication and integration of technologies already in place; only 20 percent should focus on new technology. She notes that one problem with forecasting is that there are so many kinds, from revenue projections to inventory levels to the aforementioned sales targets. Various business units pursue the form most relevant to them, yet a CFO may benefit most from a "one-number" forecast, in essence an agreed-upon view of what the future will look like.
That requires considerable coordination, which is much more difficult than buying software. J. Scott Armstrong, a forecasting expert at the University of Pennsylvania's Wharton School of Business, takes Cecere's point even further. "There are 139 principles of forecasting," he says. "Businesses ignore most of them. And little of what we've learned has been built into software." His Web site, forecastingprinciples.com, takes those serious about better forecasts through and beyond the basics. Do we foresee a huge spike in site traffic? The outlook is hazy.