The impact of social referrals on consumer conversion
Abstract
Social media is especially useful for studying social learning, peer effect and word-of-mouth (WOM) referrals. In this paper, I will use panel data from private display advertising campaigns on Facebook to study the effect of socially referred ads on online consumer conversion. In my dataset, I measure consumer conversion as the number of people who become fans of a product after seeing the product’s ad on Facebook. Facebook users receive a product’s ads in two ways. One group of users receives ads served with social context. Ads served with social context will show up on the users’ Facebook page, and with a message that informs the user that one of his friends also likes the product. The other group receives ads served without social context. This group of users only sees the products’ ads, but no message about whether his friends like the product or not. The data can thus be used as a control study. I conducted a fixed effects regression analysis. My study confirms that social referrals do have a statistically significant positive effect on consumer conversion and it does significantly increase consumer conversion. Socially referred ads have a coefficient almost 8 times as much as non-socially referred ads. In addition, a unit increase in the percentage of ads served in social context raises consumer conversion rate by about 8%. A statistically significant impact of social referrals implies the existence of a social multiplier. The social multiplier can be estimated by regressing actual aggregate level conversions on aggregate predicted conversions. The social multiplier is estimated to be 1.5. The aggregate level actions are 1.5 times larger than the aggregate individual actions. For Facebook users that have an average number of 100 friends, this implies a social interaction parameter of the size 0.345. In other words, a Facebook user’s action is influenced by about one thirds of the average of his friends’ action. The finding confirms the significance of word-of-mouth referrals and gives an estimation of the effectiveness of advertising on platforms in which social interaction is strong.
JEL code: L10, M31, D12, M37