Warc, 14 July 2014
NEW YORK: Summertime is budget-setting season for many marketers, and a study in the current issue of the Journal of Advertising Research adds a new twist to this annual exercise.
Landing on an advertising budget should not simply be a number-crunching experience, contend Douglas West (King's College London), John B. Ford (Old Dominion University) and Paul W. Farris (University of Virginia Darden School of Business).
In How Corporate Cultures Drive Advertising and Promotion Budgets: Best Practices Combine Heuristics and Algorithmic Tools, they divide budgeting methods into two broad forms.
One involves using heuristics, like making choices based on market share, what a rival is spending, a percentage of absolute sales and simply was is "felt" to be necessary.
The other depends on algorithmic methods, based on contributors like quantitative models, incremental testing, objective tasks and return on investment.
A combination of both can lead to more reliable (and flexible) marketing-spend allocation, and to an end result that reflects not just the culture of the marketing department, but the entire organisation.
"The results clearly showed that the greater the participation of the marketing personnel, the more likely algorithmic methods would be used to augment, but not totally replace, heuristic methods," the authors suggest.
"What appears to be the case is that marketers preferred logic and probability; when their participation was more diluted by other functional areas, the likelihood of using heuristics increased."
And the authors support their discussion with an analogy drawn from the great American pastime.
"What the baseball player will do is run toward the ball, constantly adjusting his speed to maintain as constant an angle as possible. A series of adjustments are made as the forward progress advances. The catch is made largely because of the outfielder responding to a series of heuristic," they write.
"It is the contention in this study that managers should think of heuristic budgetary processes in similar ways and have confidence in their value in conjunction with algorithmic techniques."
Data sourced from Warc