Researchers from the University of Southern California, University of Houston, and Uber Technologies, Inc. published a new paper in the Journal of Marketing, which finds that in order to create viral ads, brands should arouse strong emotion, place brand mentions at the end of the video, keep ads to a moderate length of 1.0 to 1.5 minutes, and use authentic characters. To arouse emotions, a brand should create an ad with a captivating plot, a surprising ending, and authentic characters; they also should use babies and animals more than celebrities.

The study forthcoming in the July issue of the Journal of Marketing titled "The Critical Role of Information, Emotion, and Brand Prominence," is authored by Gerard J. Tellis, Deborah J. MacInnis, Seshadri Tirunillai, and Yanwei (Wayne) Zhang.

YouTube is a media channel where millions of users create and share billions of videos without charge. It has also become a key platform for advertisers. Brands value YouTube because of the opportunity to reach more than one billion unique users who watch more than one billion hours of video daily. YouTube provides a low-cost and flexible platform for sharing ads with a path to wide viewership if an ad goes viral.

The research team tested five hypotheses about what drives sharing of video ads across social media, using two independent field studies that analyzed 11 measures of emotion and over 60 ad characteristics. The study included 109 brands that were among the top 100 US advertisers in 2012 as well as additional brands that were historically active on YouTube.

Key findings include:

• Of the ads studied, 10% were not shared at all and more than 50% were shared fewer than 158 times.
• Information appeals have a strong negative effect on sharing except when the advertised item involves risky purchase contexts such as new or high-priced products.
• Ads that evoke positive emotions of inspiration, warmth, amusement, and excitement stimulate strong positive sharing. Despite this fact, only 7% of YouTube ads studied evoked positive emotions. An example is Budweiser's Puppy Love ad that portrays a warm drama of a puppy that falls in love with a Clydesdale horse, is separated by the owner, and is rescued by the horse.
• Ads that use drama, plot, surprise, and characters (such as celebrities, babies, and animals) evoke emotions and induce sharing. Yet only 11% of ads studied used strong drama and only 10% evoked surprise. In addition, 26% of ads featured celebrities, but only 3% used babies and animals, even though the latter are more effective at driving shares.
• Prominent brand placement impairs sharing: Lengthy, early, or intermittent placement of the brand name drives less sharing than late placement. Surprisingly, only 30% of the ads studied used late placement.
• Ads between 1.2 to 1.7 minutes in length are the most shared. However, among the ads we shared, only 25% were between 1.0 and 1.5 minutes. Fifty percent of the ads were shorter than a minute and about 25% were longer than 2.0 minutes.
• Emotional ads are shared more on general platforms (Facebook, Google+ (which shut down in August 2019), and Twitter) than on LinkedIn. The reverse holds for informational ads.

"Our findings provide marketing and media managers, advertisers, and copywriters with specific theory-based insights into how to design ads to drive virality," says Tellis. "While the old mantra touted exposure, exposure, exposure for brand names, we find that minimal brand exposure, discreet information, and strong emotion are key drivers of virality."