Live from InterACT: How to Calculate Your Decision Yield
(Posted by Guest Blogger, Ian Turvill. Given that I have so far posted nothing, and James has been posting like a madman, he pointed out to me last night that I was lagging dramatically behind. So here's a small step in redressing that balance, though I suspect I'll get nowhere close to catching up.)
Like James, I am attending InterACT San Francisco 2007 this week and blogging live (or nearly live) from the sessions. This session is "How to Calculate Your Decision Yield", and John Nash and Joe LaLuzerne from Fair Isaac's Corporate Strategy and EDM Methodology team explored different frameworks for exploring Enterprise Decision Management, the ways in which it can be applied across organizations, and the benefits that it can bring.
The concept of Decision Yield is one that has been explored on this blog ad nauseum. In fact, we have a whole category dedicated to it, so I won't repeat those elements of John and Joe's presentation that are reflected there. Instead, I'll focus on some new approaches to thinking about EDM that haven't seen the light of day before.
Many of these new perspectives are a direct result of this week's greater focus on explaining the tangible and concrete elements of Enterprise Decision Management, or "Making EDM Real", as we've been saying around here.
Influence Diagrams Help Drive Creation of Decision Performance Audits
One of the hardest things in calculating Decision Yield is creating the survey tool that serves as the basis for calculating Decision Yield in the first place - what we have called in the past a Decision Performance Audit. To this end, Joe and John introduced the idea of an "Influence Diagram" that links the functional areas of a business under consideration, through the range of specific, quantifiable metrics which together make up the measurement of Decision Yield, down to the specific decision-making capabilities the business needs to succeed. I have attached some slides below to illustrate, but this concept migh be best explained through a specific example.
Think of the "Customer Management" function within a typical credit operation:
- The efficiency and effectiveness of decision making can be measured through a range of financial and operational metrics which can be combined to assess overall Decision Yield. These include financial metrics, such as the level of bad debt write-offs and the cost per account collected, and operational metrics, such as the number of managers and analysts required to oversee the portfolios.
- These business metrics are in turn driven by the nature of the decisions made in a number of areas, such as setting credit line increases and decreases or determining terms, conditions, and features of new product offers.
- And finally the decisions rest on certain decision-making capabilities such as the ability to properly assess risk (say, through the FICO score) or to evolve decision-making strategies through the application of Adaptive Control
EDM Adoption Continuum Helps Map High-Level Decision Yield Performance
A further concept introduced this week at InterACT was the EDM Continuum. As companies go further in the adoption of EDM, they typically pass through three steps:
- Automate: Automate decisions through rules management software
- Improve: Improve the quality of decision-making through predictive analytics, decision models, and adaptive control
- Connect: Connect decisions across the organization through reusable and open components and shared decision capabilities
At a rather approximate level, it might be said that a company which has only automated its decisions has a low Decision Yield, while companies with automated, improved, and connected decisions have high Decision Yield.
Using the lens of the EDM Continuum allows us to define best practices as a company's approach to decision-making matures and specifically to think about the value to them of thinking about this step. Consider practices in insurance underwriting:
- An insurer beginning to adopt EDM, with significant numbers of business policies, operating rules and/or regulatory requirements, will apply a Business Rules Management System to capture them and leverage them across multiple business processes. They are aiming to make their enterprise more agile and to make more consistent decisions.
- By contrast, an insurer with trying to attain the highest Decision Yield rating will be seeking to systematically advance customer value and loyalty across all transactions
Again, consult the slides below to see the EDM continuum in greater detail.
You can find the full set of posts from InterACT in this category.