While one of the primary roles of advertising has always been to build long-term brand capital for a business, the emergence of digital media―including social networks―has led marketers to focus more on response, whether that’s a “Like,” a click to a website, or a purchase.
Earlier this year, I co-authored a research paper with Wharton PhD student Dokyun Lee, along with Professor Harikesh Nair from the Stanford Graduate School of Business. The paper, entitled “The Effect of Advertising Content on Consumer Engagement: Evidence from Facebook,” investigates the effect of advertising content on customer engagement.
More specifically, we studied the role of advertising content in driving consumer engagement on Facebook Pages, while documenting the content that matters the most, whether it was informative, persuasive, or a combination of both. Why Facebook? Pretty simple. Facebook has no limit on content length and uses real names. We also considered differences in industry type, while analyzing 100,000 messages posted by a panel of about U.S.-based 800 firms over an 11-month period.
The 2013 Chief Marketing Officer Survey reports that spending on social media marketing will grow from 8.4% of firms’ total marketing budgets to about 22% in the next five years. But the debate continues regarding the effectiveness of advertising on social networks. In fact, one study cited in our research revealed that only about 1% of an average company’s Facebook fans engage with the brand. At the same time, executives at companies like TripAdvisor are rethinking the value of Facebook advertising.
My co-authors and I developed algorithms to analyze messages both linguistically and semantically. We used two metrics for user engagement: Likes and comments on posts from users with complete demographic data.
One of our biggest challenges was dealing with EdgeRank, Facebook’s filtering algorithm. If we observed that certain posts (say photos) got a lot of Likes, we needed to be able to determine whether this was because consumers engaged with those posts or Facebook’s EdgeRank algorithm caused those posts to be shown to the most active users. Obviously, the implications of the two findings are very different. We addressed this by building a model of Facebook’s EdgeRank algorithm and then analyzed consumer response after accounting for Facebook’s decision.
I invite you to read the entire research paper, but our analysis revealed that persuasive content―such as emotional and philanthropic messaging―has a more positive impact on engagement. This suggests that companies benefit from sharing their brand personality and information about their social initiatives rather than product information.
Conversely, content that included prices, availability, and product features reduced engagement when they were included in isolation, but increased engagement when they were combined with persuasive content, further emphasizing the need for companies to adopt more emotional messaging across their marketing campaigns.
This is not to suggest that firms should eschew product informative messages, especially given Likes for messages with information about products and promotions may be more valuable than Likes for other messages. Instead, it suggests that the ideal message mix should seek to balance the two and ensure that the desire to share product and brand information does not take over one’s social media strategy.
As companies increase their activity on social media, the role of content engineering will continue to grow in importance. Content engineering seeks to develop advertising content that better engages targeted users and drives the desired goals of the marketer. Companies that will succeed in today’s environment are ones that place close attention to their content engineering strategies.

Kartik Hosanagar is Monetate’s chief scientific advisor. He is responsible for data integrity of Monetate’s statistical analysis, and advises the company’s data teams in the use of mathematics and optimization. He is also a tenured Associate Professor of Internet Commerce at The Wharton School of the University of Pennsylvania. He is on Twitter as @enabletech.