Marketing researchers and marketers who are responsible for forecasting data are frequently faced with a difficult decision: finding models that will produce the best forecasts. With Forecast Pro, a statistical forecasting package designed for PCs, such a decision is no longer difficult.


A product of Business Forecast Systems, Inc., Belmont, Mass., Forecast Pro is an advanced forecasting package that allows the business user to prepare forecasts of time series data quickly, easily and accurately. It includes three basic methodologies - Exponential Smoothing, Box Jenkins and Dynamic Regression - that will solve nearly any business forecasting problem.

In Forecast Pro, form follows function. Its modular form - data handling, model building, diagnostics, forecasting and utilities - follows the function of the forecasting process. The expert system framework guides the user to the optimal technique and model for one's particular forecasting application.

The program can be used in two different modes: interactive and batch. In the interactive mode, the user constructs a forecasting model interactively under the guidance of a built-in expert system or the user's own. The batch mode lets the user develop a command file based on interactive use of Forecast Pro. Forecast Pro can then play back the user's commands without the need for human interaction.

Owens Corning

Forecast Pro is widely used at Owens Corning Fiberglas, says Bruce Harned, market development manager in the Toledo, Ohio, company's roofing products division.

According to Harned, the roofing products division has two primary needs for Forecast Pro: to find out what the general size of the market will be for roof area on an annual basis in the residential roof and flat roofing markets. Residential roofs use shingles, and flat roofing, while found in residential roofing, is primarily found in the commercial market. "We want to find out what the total opportunity is to sell in these two markets on a national, regional, and district level basis."

Forecasting is also performed to project the market share for particular products the department uses. "For example, with shingles, we examine the varying growth rates and the variables that determine market share."

In a simplified fashion, Harned explains how Forecast Pro works. "We plug in several variables such as economic conditions in specific geographic markets. Then we plug in product sales. Forecast Pro then comes up with the best methodology to forecast and the most significant variables. If you accept the recommendations for the model it proposes, it runs right into the forecasting option."

Harned praises the system's easy-to-use format. "It's user-friendly. It provides a simplified way of doing what can be a highly complex process. No service of which I am aware projects roofing opportunities as simply as ForecastPro."

FTD

Cost was initially a major reason why Florists' Transworld Delivery (FTD) Southfield, Mich., turned to Forecast Pro. Harriet Gallu, project supervisor of quantitative research in FTD's research department, says it was very expensive going through an outside vendor to forecast FTD's data. "We were also limited by the vendor's schedule and received updates only twice a year. That's when we decided to check out different software packages," says Gallu.

"In general, the system gives us a working forecast," notes Gallu. "By making adjustments for seasonality and changing market conditions, we can improve the accuracy of the forecast." Gallu is confident about the accuracy of the numbers they receive because they have tested the forecast with actual data from two to five years ago and projected last year's levels with Forecast Pro.

Having control over the forecasting system is a big advantage to FTD's research department. "We can manipulate data on a monthly basis and change environmental variables as information becomes available or as we see appropriate. Plus, even though I don't have a strong background in forecasting, Forecast Pro is easy to follow."