Viral marketing meets social advertising

摘要: Social advertisement is one of the fastest growing sectors in digital landscape: ads form promoted posts are shown feed users a social networking platform, along with normal posts; if user clicks on post, host (social network owner) paid fixed amount from advertiser. In this context, allocating to typically performed by maximizing click-through-rate, i.e., likelihood that will click ad. However, simple strategy fails leverage fact can propagate virally through network, endorsing their followers.In paper, we study problem viral-marketing lenses. We show allocation takes into account propensity for viral propagation achieve significantly better performance. uncontrolled virality could be undesirable as it creates room exploitation advertisers: hoping tap virality, an advertiser might declare lower budget its marketing campaign, aiming at same large outcome smaller cost.This challenging trade-off: hand, aims leveraging and effect improve advertising efficacy, while other hand wants avoid giving away free service due virality. formalize ad minimum regret, which NP-hard inapproximable w.r.t. any factor. devise algorithm provides approximation guarantees total all advertisers. develop scalable version our algorithm, extensively test four real-world data sets, confirming delivers high quality solutions, scalable, outperforms several natural baselines.