Virality, Network Effects and Advertising

Many video ads are designed to go viral, so that their dissemination depends on customers sharing the ads with their friends. This paper explores whether there is a trade-o between achieving this virality and the eectiveness of the ad at persuading a consumer to purchase or adopt a favorable attitude towards a product. In other words, do ads, by being the kind of ads that achieve virality, sacrice elements that would be better at persuading people to actually buy products? The analysis combines data on the real-life virality of 400 video ad campaigns, and crowd-sourced measurement of advertising eectiveness among 24,000 consumers. Eectiveness is measured by randomly exposing half of these consumers to a video ad and half to a placebo ad, and then surveying their attitudes towards the product. We nd that ads that were more `viral,' that is, ads that had achieved more views on websites such as Youtube.com, were indeed less eective at persuading consumers to purchase or adopt a favorable attitude to a brand. Relative ad-eectiveness dropped by roughly 10% for every million views. Taking into account the advantages of increased reach, this means that there was a decline in overall advertising eectiveness at 3-4 million views. Importantly, ads that generated both views and consumer engagement in the form of comments did not suer from the same tradeo. Such ads were also be less intentionally provocative or outrageous than ads and more likely to be viral due to humor or attractive visual-design