This paper brief reviews some evidence often quoted in relation to quantitatively evaluating advertising expenditure. The need to relate quantitative methods to facts and theories of consumer behaviour is emphasised. A view of consumer buying behaviour is discussed and related to three possible kinds of advertising effects which are distinguished; immediate, medium-term and long-term effects. It is suggested that medium- and long-term effects may be intermittent and questions are raised as to whether they are currently able to be measured by econometric analysis. A recent hypotheses that long-term effects, even if ignored, are unimportant is discussed and questioned. It is finally proposed that some important aspects of advertising strategy are best approached via area experiment. A longitudinal approach to studying the role of advertising in a brand's development is suggested to cover the investigation of the effects of different levels of advertising early in the development of advertising and for very long periods of time.