%0 Thesis %A De Canha, Nicholas Manuel %D 2017 %T On the thresholds and ceilings of advertising effects moderated by clutter and familiarity %U https://bridges.monash.edu/articles/thesis/On_the_thresholds_and_ceilings_of_advertising_effects_moderated_by_clutter_and_familiarity/4697440 %R 10.4225/03/58b3bd01504af %2 https://bridges.monash.edu/ndownloader/files/7664017 %K Ceiling %K Decay %K thesis(doctorate) %K Goodwill %K Data envelopment %K Familiarity %K 1959.1/1204721 %K Efficiency %K ethesis-20150618-16021 %K Herfindahl %K Clutter %K Advertising %K monash:160302 %K Threshold %K 2015 %K Open access %X The effect of advertising on sales and market share is complex, as are the interactions between advertising and the environment in which it occurs. This study investigates the effect of advertising on sales, including providing unique contributions by quantifying threshold and ceiling levels of advertising effectiveness and providing an understanding of what affects those levels. Specifically, the study considers the effects of familiarity, market share and clutter on advertising effectiveness, and on those thresholds and ceilings, and quantifies the relative importance of each of these factors by combining them into a single model. The determinants of advertising goodwill decay are also considered. The study uses two methodologies: Data Envelopment Analysis and nonlinear regression, on a longitudinal (eight-year) data set of over 6,000 observations. Findings confirm the existence of ceiling and threshold effects whilst demonstrating the impact of clutter and familiarity. The share-of-voice / share-of-market relationship is re-examined in this context. The findings further demonstrate that the ceilings and thresholds are dependent on the level of clutter and familiarity. These findings extend the existing literature by providing evidence from the field (as opposed to previous laboratory-only studies), and by including several new variables (the Herfindahl Index, a measure of advertising concentration, and the difference between Share-of-Voice and market share) together with variables of clutter, familiarity and lag effects, simultaneously, in a single estimator function. %I Monash University