Research Brief: Do Multi-media Advertising Campaigns Pay Off?

In the era of Big Data, marketing managers must move beyond analyzing aggregates of market data


What’s the best way to stretch the advertising budget for the newest car across an array of media platforms? The conventional way — by mining heaps of market data from all media sources — can be misleading and inefficient. Standard data collection methods also fail to capture the complexity and diversity of multi-media effects; after all, consumers react differently to television advertisements than to print ads or web banners. Marketing managers need to work out how these various media prompt varied responses, and how media platforms work together or even against one other. This study presents a new method of advertising analysis that appears to do exactly that. 


The researchers used the MARS (Multivariate Adaptive Regression Splines) modeling approach, first developed in 1991, to analyze the multi-media effects of more than 50 brands (SUVs, hybrid cars, and beer). They took into account a large number of media effects, both on a stand-alone basis and interacting with other media. They then identified the threshold (when consumers started to notice advertising) and saturation points beyond which advertising spending stopped having a positive impact. 


  • The study empirically confirms the famous assertion that “half my advertising is wasted.” A full 55 percent of all advertising investment was found to be unprofitable.
  • Showing an advertisement across as many media platforms as possible can have an undesirable effect. This multi-media bombardment can actually trigger “interaction super-saturation” — ads become boring or even annoying. Strategic moderation is key.
  • The threshold and saturation points vary for every product and medium, but spending outside of these ranges does not benefit the advertiser. For instance, a network TV ad for a Ford Explorer started to be profitable after 20 percent of allocated spending, but was saturated at 55 percent (65 percent of the spending on the ad was inefficient). The same vehicle, advertised in a newspaper, maxed out at a mere 5 percent, meaning that 95 percent of the spending on the ad was inefficient.
  •  MARS comes out with flying colours as an ideal methodology to capture the complexity of multi-media advertising: “It performs well even for a large number of media, it allows for easy calculation of turning points (threshold, saturation, and super-saturation) and detection of inefficient investment allocation, and it can be easily implemented using available software.”


Marketing managers must take advantage of rich new methodologies, such as MARS, to move beyond analyzing aggregates of market data. Looking at aggregate response usually tells very little about unique consumer reactions to single forms of media or the interactions among multiple media; worse, it offers no useful guidance on how to avoid inefficient investment. In the era of Big Data and proven methodologies, it is now easier than ever to pinpoint where advertising dollars are being spent profitably or wasted, which is just as well, since so many of advertising dollars miss the mark. “An immediate priority of future research,” say the study’s authors, “should be the study of optimal media allocation decisions acknowledging both the wealth of the media choices and their congruence within the firms’ communications mix.”

Title: The complexity of multi-media effects 

AuthorsCeren Kolsarici (Queen’s School of Business) and Demetrios Vakratsas (McGill University)

Published: Marketing Science Institute Working Paper Series 2011 (Report No. 11-100). Working paper version available on the Marketing Science Institute website.

Nick Walker

Smith School of Business
Goodes Hall, Queen's University
Kingston, Ontario
Canada K7L 3N6

Follow Smith Business Insight on: