|
|
|
Among the more common problems is the inclusion of the amount spent on advertising with the amount invested in price promotions. Often changes in price will prove to be related to sales but changes in the amount spent on advertising won't. So the model that is constructed shows price has a substantial effect, but advertising doesn't. Actually, if the amount spent on advertising had been modified to reflect the impact, or quality, of the advertising it would have been related to sales. The true importance of advertising would have been revealed, and the model would have predicted sales more accurately. Don Bruzzone was first asked to expand on these views at the IIR Marketing Mix Conference in October 1999. The talk "Improving Marketing Mix Models With Data on Ad Quality" is one of a number on the subject listed in section on Talks and Articles. He chaired the IIR Marketing Mix Conference in March 2002, and presented a paper that has since been published in the March 2003 issue of Quirks: "Modeling Advertising? First determine if it was good, bad or average." It presented further evidence, based on our years of tracking Super Bowl commercials, that the top 20% reaches and affects eight (8) times as many people as the bottom 20% - for the same amount of money. So, if your model only includes the amount of money spent on advertising, there is little hope it will show the true effect your advertising has on sales. BRC develops marketing mix models, and it consults on marketing mix models developed by others to be sure the true effect of advertising is being reflected as accurately as possible.
|