a young female student using a cellphone and credit card outside on campus

Selling through Social Networks: Impact of Social Ties and Product Characteristics

Since 72% of Instagram users make purchase decisions after seeing something on Instagram and over 40% of respondents in a recent survey made online purchase after seeing it used by an influencer on social media platforms like Instagram, Twitter, Vine and YouTube, product exposure and connectivity on social networks positively impacts online sales. 

Liangfei Qiu, Arunima Chhikara and Asoo Vakharia

From left: Liangfei Qiu, Arunima Chhikara and Asoo Vakharia

Using an analytical framework and experimental evidence in an observational learning setting, recently completed research from Warrington College of Business Associate Professor Liangfei Qiu, Ph.D. student Arunima Chhikara and McClatchy Professor Asoo Vakharia offers key insights into how social ties and product characteristics moderate online purchase decisions. A key finding is that prior purchase experiences of any social network user (friend or stranger) positively impact future sales for vertically differentiated products (e.g., ISP providing faster internet access). 

However, this is not the case for horizontally differentiated products (e.g., bottled water brands which are preferred based on individual taste) where only the prior purchase behavior of friends positively impacts future sales. More nuanced is the finding that contact-based ties rather than structural ties enhance the impact of observational learning. 

From a practical perspective, implications are that online retailers should adopt different strategies for enhancing sales. On the one hand, for quality-driven products, incentivizing all prior purchasers would be the preferred option for an online retailer. Conversely, for products where consumers preferences are driven by taste, then targeting influencers with large networks to recommend their product is a better strategy. 

This research will be published in Information Systems Research.