Rolling forecast model can improve usefulness of financial data

At the Boston IBM Finance Forum, one expert urged finance professionals to adopt a rolling forecast model to replace slow and broken processes.

How's this for sound advice: "Stop doing dumb stuff."

So said Steve Player, guest keynote speaker at the Boston IBM Finance Forum held at the New England Aquarium. Player, founder and managing partner of The Player Group based in Dallas, asserted that current forecasting techniques are broken, and urged the audience of approximately 60 finance professionals to move to a rolling forecast model.

According to Player, many finance departments today are positioned at the "back of the ship" -- measuring and analyzing past performance rather than helping to anticipate and prepare for future events. "If you take a hard look at it, [finance] almost exclusively looks at lagging indicators," he said. But with the time that's freed up by discontinuing "dumb" processes, Player said finance can take on a more predictive role that could help organizations stay vigilant in uncertain economic times.

Finance should move away from 'batch mode'

Player described current forecasting practices as "forecasting to the wall." While most attendees indicated that they produce year-long forecasts, Player said forecasts often get increasingly shorter-reaching as the year progresses. For example, three months into the year, the forecast will be recalculated for nine months, and six months in, it will be recast for the next six months. This pattern continues until "somewhere down there in the darkness you forecast another 12 months," Player said.

"If that's your forecasting system, the view you have is distorted," he continued. "That's what masquerades as forecasting at most companies today -- it's not about visibility, it's about a revalidation of a performance target."

Player also pointed to one aim of finance executives that might not be entirely worthwhile: faster monthly closes. "Finance is in batch mode, [but] the world doesn't live in batch mode anymore," he said.

He proposed a hypothetical situation where it was possible to instantaneously close the books at the push of a button on the last day of the month. "Would your operating people have a midnight review session of that hot financial data? What could they do [besides] validate thousands of decisions they've already made? There are a lot of good reasons to close the books faster, [but] if you say that it's to give operating people more timely information, it's a bogus reason because even if I could do it at the stroke of midnight on the last day of the month, it's too slow," he said.

Player said this example points out the need for a more continuous data model. "If we in finance are going become relevant, we've got to get off this batch mode," he said. "What do we use all this big data for if we have to wait until [the] month's end? Why wait?"

Five characteristics of effective forecasting

Player went on to explain the value of a rolling forecast model, which entails producing forecasts on a continuous basis that extend for a set amount of time. While he said a rolling forecast might not always be totally accurate, its value lies in its usefulness. For example, if a forecast predicts an undesirable result several months in the future but executives are able to address the underlying factors before the result is realized, the forecast wasn't accurate, but it was extremely useful.

Player also listed five attributes of "diamond-level" forecasting: timely, actionable, reliable, aligned and cost-effective.

More on financial management software

Learn the basics of financial management software

Find out the top five reasons to trust cloud financial software

Get tips on how to choose a financial management system

How timely the forecast is depends on a company's adaptability, Player explained. "The slower you are at adapting, the longer forecast horizon you need," he said. "You can solve this problem [by] getting longer forecast horizons or getting more nimble." He also advised finance professionals to take stock of which data points are either highly critical or highly variable or both, and check up on those more frequently than others. Planning at the chart of account level, he said, often wastes time by focusing on "little rocks."

An actionable forecast advises executives what actions to take in adequate time to avoid a bad consequence. In terms of reliability, Player advised the audience not to overcompensate or undercompensate when constructing a forecast. "If you always beat your forecast, you are the definition of a biased forecaster," he said. "You need to separate forecast from target -- the forecast needs to be the unbiased look at where we really think we're going to go. The minute we put out anything to bias that, we start destroying [the forecast's] usefulness." Alignment refers to the importance of having one version of the forecast that all business units work from.

He also suggested that organizations plan for a wide set of outcomes, both good and bad, to counter confirmation bias. "Build a set of leading indicators that tell you when any of those realities are starting to come true, and then you'll be in better shape to know where you're trying to get to," he said.

And Player pointed out that this approach, in addition to implementing a rolling forecast model, is especially important in a world where many events are beyond our control. "Think about all the things in your business life that are coming at you this year that you can't stop. Can we stop anything happening in Europe from hitting us? Can we stop the fiscal cliff in our federal government? Those hurricanes are coming, and all we can do is make sure we're ready for whatever path they choose to take," he said. "That's the reason we have to get off the back of the boat and start to become forward-looking in our forecasts."

Emma Snider is the associate site editor for Follow her on Twitter: @emmajs24.

Dig Deeper on Financial analytics and reporting