James was interviewed by Madan Sheina for the July 16th issue of Computergram
Section: 04. Business Intelligence
By Madan Sheina
James Taylor is vice president of Enterprise Decision Management at Fair Isaac, an IT vendor perhaps best known for its FICO credit scoring system. However the company is also preaching an approach to tackling corporate decision making called “Enterprise Decision Management” that focuses on automation and management of decisions rather than taking action on business intelligence and analytic data.
Taylor has spent several years developing Fair Isaac’s EDM product strategy and is the company’s most vocal front-man on the subject. He recently co-authored a book with Neil Raden, founder and president of research firm Hired Brains, called Smart Enough Systems: How to Deliver Competitive Advantage by Automating Hidden Decisions. He spoke with ComputerWire about the key themes of EDM in their book and how it can smarten up organizations.
Q. In a nutshell, what’s EDM all about?
A. Decisions can, and should, be automated and managed in a formalized and disciplined way. Externalizing and managing decisions is the key; after all, customers think your business decisions are deliberate, so they should be.
Q. So who should read this book and why?
A. Everyone should read it so that Neil and I can retire [laughs]. No, but seriously, business people who care about the way their business is constrained by their information systems should read it to understand why current approaches to automating decisions don’t work. Information systems developers should also read it to understand how they could adopt new techniques and technologies to build systems that are more likely to delight their users. And finally, data people should read it to see how the corporate data they’ve been assembling, cleaning and managing can be put to business use.
Q. The title implies that today’s IT decision support systems don’t make the grade?
A. It’s not just systems for decision support that are not up to scratch. Because decision-making is considered a manual task — something done by people and, at best, supported by systems — decision making that must be embedded in systems is largely left to programmers. Other decisions are missing and not handled by systems at all resulting in needless referrals to supervisors, delays and low rates of straight-through processing.
Q. What do you mean by “missing” decisions?
A. Missing decisions are operational decisions that you likely don’t see. They could be manual decisions, taken by someone in the organization, for example: a decision to waive a fee for a cranky customer. Or they could be hidden behind a one-size-fits-all decision, like what to put on your home page or what options to use in a kiosk. Or perhaps they are implicit business decision taken by a piece of software and so on. Most organizations have far more operational decisions than they think because they fail to consider all the decisions hidden in their business. If you cannot say what all your decisions are, how do you know they are being taken appropriately?
Q. How can adopting an EDM approach help?
A. The key idea behind EDM is simple; that a focus on decisions combined with the right technology enables decisions to be improved. As one of the founders of Fair Isaac once famously said, “if you want to improve a decision you have to grab it by the throat and not let go”. If you hide it in a system or a process you will not be able to really improve it. This singular focus is essential as decisions are different. The “best” decision is also not a static thing. As competitors, markets and business needs change so must the way a decision is made. Thus an automated decision must constantly be revisited, tweaked, tested and improved. EDM is the best approach for doing that.
Q. How does that differ from traditional business intelligence and analytic tools?
A. EDM is basically different from traditional BI and analytics in its focus on taking action. EDM is about automating and managing decisions, not just helping people make them. BI is about making sure that a person has the data they need to make a good decision. Hence EDM goes a lot further and makes sure a system knows how to make a good decision. It’s also true that traditional BI and analytics is rear-view and tends to be focused on analyzing the past, whereas EDM is focused on changing the future.
Q. What are EDM’s core tenets?
A. There are really four parts to EDM. The first is focusing on decisions as distinct opportunities for improvement and drivers for automation, typically by designing and building decision services that deliver decisions to other services and applications within the application architecture. Then there is the application of business rules, to provide a decision automation platform that ensures consistency and delivers agility by allowing the business to collaborate with IT to make rapid, accurate changes to the way the business decides, and thus acts. Analytics, especially predictive analytics, is important to leverage the data you have by deriving insight from that data and using that insight to improve those decisions. And last but not least you need adaptive control to continually test and refine your approach to each decision to find better and better ways to make the decision and to ensure that changing circumstances have not undermined the value of your decision.
Q. So how many organizations out there are smart enough today?
A. In my experience very few organizations are, though individual areas of many are. For example, many insurance companies have a smart enough underwriting system. Many banks manage their credit risk effectively. Some marketing and loyalty programs are smart enough. In general, few if any companies have an overall application architecture that is smart enough and many companies have a whole set of applications not one of which is really smart enough for the world as it is today.
Q. Does that mean they’re not ready for EDM?
A. Readiness is typically a combination of several things. The first is better collaboration between IT and business on systems; if they can’t get on you need to find out why. Then there is data readiness, for example does the organization have the data it needs, can it access it, is it accurate? Companies also need to demonstrate a level of analytic understanding to be able to derive insights and understanding from its data. On the softer side of the equation the company also needs to be willing to change and adapt to new technologies and approaches and management needs to focus on operations; for example, do they regard execution and operational excellence as worthy of their time or are they all obsessed with strategy and planning? But having all in place beforehand isn’t a mandatory prerequisite for EDM. Companies can get started no matter where you fall on these measures of readiness. But they do need to understand and allow for the reality of their situation.
Q. Can decision making really be fully automated? What about the human element?
A. Some decisions can be 100% automated and some must be — for instance, any decision needed by your Website, interactive voice response (IVR) system, or automated teller machine (ATM). Many other decisions can be 90-95% automated with only unusual cases needing manual intervention. Many of those factor in the human touch by giving choices or including judgmental factors such as customer mood.
Remember, though, you’re not really factoring out human judgment. You’re just changing which human is making the judgment. Instead of the human with the interaction, usually underpaid and over worked front-line staff, you use the judgment of an analyst who can see what the data is saying and an expert who understands the regulations, the policies and who has the experience to make judgment calls. Organizations should assess who is actually taking a given decision and then decide if they really want that person to make that decision. For example, do you want your call center representative deciding if a fee can be waived? Or do you want a programmer deciding which options to display for a customer? EDM lets you take full control of these decisions.
Q. What do you see as being the market opportunities for EDM?
A. Today EDM is mostly widely applied in what you might call risk-centric decisions like credit, insurance underwriting, claims, fraud, and so on. Clearly those are areas are the low hanging fruit right now. The use of EDM to manage customer experience decisions is the short term area of growth across the board. In the long term I think the use of EDM to deal with new data sources from radio frequency identification, mobile devices, remote monitoring, and so on will be big.
Reproduced with permission from ComputerWire’s Computergram (subscription required)