Skip to main content
Automating and improving pricing in banking

I saw this piece on "Best Practices in Customer Management: Some New Methods Breaking Out" by Kathleen Khirallah over at Tower Group. Kathleen is always a thoughtful writer and this piece was no exception. One of the take-aways struck me particularly:

"If there is one factor that has hampered banks' ability to be customer-centric and proactively manage their relationships with customers, it is their reliance on a "one-size-fits-all" approach to the mass market."
Now Kathleen talks about this in the context of pricing for this paper, but I think this is a valid criticism of most banks about almost every aspect of how their interact with customers. But let's stick to pricing. Kathleen discusses why banks find it so hard to compete with pricing or even to use pricing as part of an overall strategy.

She says:

"Banks' disinclination to compete on price is generally tied directly to the paucity of analytics and rigor in their pricing computations".

Essentially pricing can only get more sophisticated as more analytics enter the decisioning process. Simple segmentation analytics will help but any serious attempt to manage pricing in a more sophisticated way will involve multiple models (risk, propensity to buy, propensity to use credit, retention risk) and do some tradeoff between them. In addition pricing decisions will still need rules as there are layers of regulation and policies that must be applied around the models. The key element to get started is that Banks need more finely grained segmentation for their pricing. Most of them already do a great job of segmentation for risk, credit line management and so on but they lack this approach in pricing. They don't have a comprehensive pricing strategy that reflects the sensitivities and desires of customers. As Kathleen says:

"If there is one factor that has hampered banks in their ability to be customer-centric and proactively manage their relationships with customers, it is their reliance on a one-size-fits-all approach to the mass market".

So how would you tackle this from an Enterprise Decision Management approach:

  • Focus on the pricing decision made for a product and a customer as a specific operational decision As distinct from saying the decision is a strategic one as to how to price a product line.
  • Build analytic models for various aspects of the customer
    • Propensity to buy
    • Price sensitivity
    • Lifetime value
    • Credit Risk
    • Retention Risk
    • ...
  • Build some kind of decision model to show how these aspects interact and are constrained
  • Offline, optimize the decisions based on this model to come up with the best rules for pricing for each customer segment Do some what-if analysis and flex your constraints to see what the impact is of these changes. Come up with the best set of rules for pricing based on this modeling.
  • Deploy these pricing rules into all the systems that need them, ideally using centralized decision management

Fair Isaac approaches this with a product we call Decision Optimizer and an associated methodology called Strategy Science. To be fair this has not been widely used in pricing yet but it has worked well in credit line management and fraud referral strategy design. Another company taking a similar approach is Earnix who provide a service to offer the best or optimal price when a customer (new or existing) asks for a product price based on this customer’s predicted price sensitivity and propensity to buy and then generate a rate card (rules for rating essentially) that can be deployed.

Finally one last comment from Kathleen:

"A few forward-thinking banks have recognized the importance of customer satisfaction and now report and measure satisfaction with the same rigor that is typically associated with risk management"

I think this reveals a key point - that even customer management decisions have an element of risk. There is a risk implicit in using resources on this customer that could be used on that one. There is a risk in a price for a product in that it might retain unwanted customers or deter potentially profitable ones. Banks have long taken a fine-grained and analytically-rich approach to risk management. Time to do the same for all aspects of customer interaction.

Technorati Tags: , , , , , , ,

related posts