As we’ve discussed previously on this blog, when we talk about customer analytics, it’s easy to think of retailers. Stores want to dissect who their customers are and what they do, both of which revolve around patterns. Demographics, web site usage, and product buying all allow businesses to predict future customer behavior.
This, in turn, gives businesses the opportunity to develop products and services to suit specific customer segments. And then marketing can better “nano-target” to the customers with a high propensity to buy.
However, it’s not just the supermarkets and clothing stores that can benefit from improving their customer analytics. As a solution architect in financial services, I see the industry transforming from product-centric to customer-centric. Financial services customer data is vast, spanning across multiple touch points such as ATMs, web logs, and phone conversations with support, but this data is typically silo’ed across an institution’s products.
Bringing all customer data across products and enriching it with data from external sources could yield game-changing discoveries. As I was working with a large retail brokerage firm and their multi-terabytes of customer data, we identified similar customers through our unique community detection system. We defined similarity by specific attributes, including account balance, portfolio mix, trading behavior and net worth. Then we could identify highly profitable customers; for example, customers trading in derivatives, futures, and the like are typically more profitable due to higher transaction costs.
This type of extensive clustering, although possible using traditional analytics methods, can be achieved more rapidly and easily with graph technology. Graphs are all about connecting entities together through relationships, so you can better discover previously unknown patterns.
To give you a deeper understanding of graph analytics, this image above is a screenshot of a bipartite graph. Each node represents a specific asset; the colors represent groups, or the output of a community detection algorithm. Assets in the same group are more likely to be held by the same institution.
Using graph technology reveals these connections much faster and more intuitively than relational databases. And once you discover the secret sauce of your customer base, you can vastly improve your marketing efforts through campaigns to cross-sell/up-sell, and thereby improve loyalty and prevent churn.
Whether you’re working at a brokerage firm or the big box store down the street, if your technology only reveals a partial view of your target audience, you’re missing huge pieces of the analytics puzzle. Financial institutions know risk better than anyone, and if they’re realizing that this is too big of a risk to take, why wouldn’t your business feel the same?