Some companies, such as heinz, can forecast revenues well using pure time series analysis (that is, by extrapolation of prior data, accounting for seasonal effects). other companies, such as fedex (which makes money by shipping packages), or sony (which sells consumer electronics) find that they cannot rely on pure time series analysis for reliable forecasting. they are strongly affected by recessions and need to use theory-based methods, including such explanatory variables as income in their forecasting models. why are they different from heinz?