Forecasting Methodology

The methodology behind IBISWorld’s data forecasts for Industry Research Reports.

Updated over a week ago

What is IBISWorld’s forecasting methodology?

Our forecasts combine analyst knowledge of industry operating conditions – competition, barriers to entry, life cycle, etc. – with their knowledge of future movement of key external drivers. Our analysts use two primary forecasting methods.

Method 1: Regression analysis

The first forecasting method utilizes regression analysis techniques. Analysts specify the models using a mixture of economic theory of industry connections, statistical testing and best practices for industry sectors developed over time.

While models vary in complexity, the selection of predictors and weightings built into the models are always supported by accuracy testing and analyst gut checks.

Method 2: Driver growth rate models

The second method of forecasting industry growth patterns uses the growth rates of key external drivers from our internal Business Environment Database. This method is effective when analysts are working with restricted data sets and volatile trends, which make regression models ineffective predictors of industry performance.

Analysts use their industry expertise and frequent updates of forecasts to understand how growth in key drivers relates to industry revenue. The growth rates of key drivers are weighted and applied to the previous year’s data to generate a year-ahead prediction. Analysts then optimize the correlation between model predictions and actual values by adjusting weights. The resulting forecast effectively models volatile and idiosyncratic industry change.

How accurate is IBISWorld’s forecasting?

We have a set of checks in place to ensure the accuracy of our forecasts.

First, we maintain a rigorous quality control process. Each portion of an analyst’s research is checked. Forecasts are held to quality standards and assumptions must be defended by the analyst.

The standard for measuring model accuracy for regression-based models (Method 1) is mean absolute percent error (MAPE). Correlation coefficients are checked for driver growth rate models (Method 2). The standard for results varies based on industry revenue volatility.

Quality checks also act as a second opinion on forecasted revenue trends. The forecast is examined in the context of previous trends, past iterations of a report and its deviation from the last year of in-sample data. As a result, quality checks result in consensus agreement on the forecast.

For any additional questions regarding IBISWorld’s forecasting methodology, please reach out to your Client Relationship Manager. If you’re not an IBISWorld member, please contact us to learn more about our membership options.

Did this answer your question?