1. Causal modeling is likely to be the best approach, because factors other than time—such as
2. Time series forecasting techniques are not terribly well-suited to developing forecasts for
multiple periods into the future because they have built-in mechanisms to incorporate past
3. The advantage of having computer-based forecasting packages lies primarily in the fact that
they can quickly develop and evaluate forecasts for thousands of products. These packages
4. Linear regression to develop a time series forecast is different from a causal linear regression
5. Forecasting is very important for firms because they depend on forecasts as an input to many
planning activities. But firms also need to address the organizational issues surrounding
forecasting—for example, how will we share data and who will be responsible for generating