Improving Forecast Accuracy For New Products With Heuristic Models

When forecasting demand for a new product, we generally use cause-and-effect models but, owing to lack of data, forecast accuracy is generally poor, especially for brand new products that have no like items. In this article, I propose using a new heuristic model for new product forecasting. I argue that, though use cases are limited, forecast accuracy of this model may be significantly higher than that of analytic models. I will also reveal how we can generate range forecasts which changes uncertainty to risk which can subsequently be managed by supply chain management best practices.