Should You Change Where Models are Deployed?
In a previous blog we noted a customer’s struggles to effectively manage the value-generation chain that takes raw data, normalizes it, creates attributes, and applies analytics in support of their business. Additionally, we’re also hearing from customers that previous decisions about the management of the value-generation chain is becoming a problem. Let’s look at two scenarios we often see in the market.
Value-generation chain within the Decisioning / Origination System
A myriad of technology platforms are used to deliver products including branch networks, loan origination platforms, mortgage systems, and commercial lending platforms. Often the attributes, models, and policies are deployed within these solutions. This practice has generally been acceptable as banks have been largely product-centric in their approach to servicing their clients.
What’s changed? Banks are becoming more customer-centric in their approach to taking products to the market.
Value-generation chain at the data provider
Some clients have leveraged analytics as a means to create competitive advantage. It is widely understood that custom analytics are generally more predictive of customer behavior than generic industry models. Typically the models themselves are built using data from a particular provider, such as a credit bureau. Consequently a bank may often choose to deploy these models at the data provider. This practice has generally been acceptable as credit data itself has largely been sufficient for segmenting and decisioning.
What’s changed? New data sources can very effectively augment and improve the predictiveness of models.
The table below shows these two scenarios, and some inherent limitations and problems with each.
Equifax can help your business better manage the value-generation chain allowing your business to more effectively manage changes that affect your business. If you would like more information on this solution reach out to your Equifax representative or email us.
This post was contributed by: Lee Grice.