Marketing Channel Recommendations in Banking

Many industries have applied various analytics techniques to personalize the experience of their customers as well as the offers made to them. Personalization has become the hallmark of every service and product industry even in the once monolithic banking sector. Recommendation systems are a core component of this personalization, focusing on targeting consumers with specific products. Traditional recommendation systems make use of large data sets in sparse matrices while conventional user targeting systems leverage external data sources (e.g., from social media and third parties). Our research focuses on data from a large financial institution and investigates the issue of recommendation systems for channel and product targeting. The proposed system achieves both high performance in channel targeting and acceptable performance in product targeting as illustrated through numerical results.