Don’t Let the Machines Take Over Business Forecasting

By Jeremy Morris, Associate Editor, US Daily Review.

As companies come under increasing pressure to respond quickly to change, many rely on business forecasts to see past the horizon.  Managers, however, are reluctant to entirely trust the forecasts that their companies produce.

In the latest issue of Analytics Magazine, business forecasting expert Brian Lewis, PhD and a Vice President at Vanguard Software, contends that forecast skepticism is well-founded.   Forecasts are traditionally produced by sifting through the millions of data points that make up a company’s past performance.  Business forecasts, he says, end up functioning more like business histories.  Hence, the only way to create forecasts executives can trust is to balance analysis of big data with human insight.

“Your staff knows things a computer can’t possibly uncover.  They have important insights about the market, future plans, new products, competitor shifts, etc.   When the forecasting process doesn’t allow people to collaborate, to really share what they know, everyone immediately recognizes the forecast is incomplete,” says Lewis.  “Forecasts lose further credibility when you can’t answer a simple question like, ‘Where did this number come from?’  If the forecast is based on numbers considered stale, then that forecast is of no real practical use.”

Recent forecasting software innovations allow multiple contributions, tracking, and real-time data pulls.  The changes enable managers to test different business scenarios and build contingency plans.  The greatest benefit to these changes, Lewis says, is business agility.

“Being able to produce new forecasts on-demand, the minute something happens, and knowing that the forecast reflects the insights from all your departments—that’s unprecedented responsiveness.  This is the future of business,” says Lewis.

All opinions expressed on USDR are those of the author and not necessarily those of US Daily Review.

Leave a comment

Your email address will not be published.


*