Some Thoughts On Bi And Analytics In Retail Banking

by FICO

(Posted by guest blogger, James Taylor)

Late last year the folks at Datamonitor published Business Intelligence in Retail Banking (Review Report). I finally got to read it today and it is, overall, a very interesting paper and I recommend it to those of you considering your overall BI/analytics strategy in retail banking. Here's the summary they post (free) on their website:

Datamonitor believes that an effective compliance solution requires a strategic investment in infrastructure that embraces assimilating business intelligence with advanced analytic tools for risk modeling and other key technologies, such as business process, change and content management.

BI has come a long way from its origins in the early decision support systems. It is currently in charge of turning transactional data into actionable information and delivering reports to an ever-growing number of information-hungry executives and managers across all lines of business.

The value of information decays over time, therefore the real-time ability is promoting business intelligence tools from their status as decision support applications into the domain of decision automation.

Now regular readers will know that I dislike lumping analytics in with BI, feeling that it tends to focus people on OLAP and reporting not executable analytics. You will also know that real-time and BI are words I don't like to use in the same sentence. That said, the report has some very valuable insights.

Compliance requirement is keeping focus squarely on reporting and monitoring functionality.
While the report is correct in saying that reporting and monitoring are critical for compliance, the reality is that compliance is about ensuring that you have followed the rules set our by a regulator. This in turn means that adopting a business rules approach and a business rules management system to manage the rules with which you must comply is a very effective tool especially when, as the report notes, there is a driving need for real-time straight through processing. If you automate transactions then you need a way to show that you processed them in a compliant way and business rules are perfect.
Sales and service effectiveness requires ability to predict customer requirements as well as fully understand customer relationship.
This means more, and more accurate, segmentation and understanding of customers across both automated and non-automated channels. It means moving to prediction analytics not just understanding and predictive reporting, hence EDM's focus on predictive analytics. The focus on revenue generating activities also shows the advantage of taking a "best next action" approach.
Strong management necessitates deep insight and control of banking operations.
BI must be fully aligned with business processes.
I am a strong believer in aligning analytics and decisioning with business processes, especially through the use of decision services. Not only do decision services represent a great point of automation for applying business rules, they are ideal for adding insight from analytics.
Ability to respond faster to customers, to regulators or to management generates need for real-time automation.
The report focuses on faster, more agile response driving a need for real-time monitoring. I would add that agility means closer, more real time monitoring AND a greater ability to respond. In an increasingly automated banking world this means being able to quickly change how systems work, another reason for business rules and EDM.

I had a couple of general points too:

This report, like so many, assumes that making BI more pervasive means giving more people access to the same kinds of systems. I do not. Pervasive BI means embedding analytic insight into the applications that run the business and that takes EDM not more BI.
There was a nice point made about the possibility of competitive advantage from faster decision making, something I have blogged about before.
There was a really nice graphic of how banks are progressing to data-driven customer service
The report emphasized the need for more dynamic product configuration and pricing and the value of EDM in this was illustrated for me in the CapOne Credit Card builder.
BI in risk and fraud is not really what most people call BI - we are talking about risk models, fraud detecting neural networks etc not reports.

Finally I think the report underestimates the technologies available to deliver real-time Straight Through Processing with automated decisioning. While it may be hard to move from a current BI approach to real time/end to end processing, this is not only approach. Banks can and have come to decision automation from their scoring and credit management business too - the place where EDM started.