James Taylor
VP of Product Marketing
Enterprise Decision Management
Fair Isaac


A great deal of attention this year has focused on the quality of corporate decision-making - or the lack thereof. Harvard Business Review recently devoted most of an issue to examining proper executive strategies. A more telling indictment of the operational status quo might come from the popularity of the long-running comic strip "Dilbert," in which bureaucracy and pomposity turn an office into an insane asylum of bad decisions. In the real world, there is a lengthy continuum of situational decisions reaching from upper and midlevel executives to front-line business rules that oblige action on myriad details of business - all intended to turn strategy into policy. James Taylor, vice president of product marketing for Enterprise Decision Management at Fair Isaac, works in this realm of technology, thinks about it, reads about it and blogs about it. DM Review Editorial Director Jim Ericson recently sat down with him to hash out some of the basics.

DMR: What's the distinction between decision management and decision automation? 

JT: At Fair Isaac we use the term decision management because the first time you automate a decision, it does not give you competitive edge; it is your ability to manage the information and evolve a strategy over time. Decision management involves an initial automation - taking control of how a decision is made, giving you the ability to actually determine how you're going to make these sorts of high-volume operational decisions in a day-to-day environment - and then managing that as a corporate asset and strategic weapon.

DMR: The idea of decision automation suggests a binary decision. Is it about setting a threshold and adjusting it periodically?

JT: At first glance some decisions appear to be that way: I give you a loan or I don't. In fact, there are multiple versions of the loan - there is the rate, there are different conditions. There's a phrase used in insurance that says, "There's no such thing as a bad risk, only a bad price." There is a price for any risk, which explains the trend moving from simplistic to more complex decision-making. If you're in an industry where you have to combine complex decision-making with fast response to competitors and high volumes, you're forced to automate the decision and manage the automation in a very active way.

DMR: We talk about contextual decision-making - that things need to be viewed in the context of a role or a position - but that doesn't necessarily align decisions with enterprise goals, does it?

JT: There are two ways in which you can view context. There's information-use context; things such as decision support and business intelligence need context for information. The same applies when you're trying to execute a decision in the context of a particular transaction. But you want that context to be driven by corporate strategy as much as possible, and it quickly gets thorny. We discovered, for example, that telcos could boost their profit margin by retaining fewer less-profitable customers. No one would do that because retention rate is about growth, which is more important than profitability when it comes to share price. Making more money would lower your stock price because retention would be down. Unless you have that kind of context - unless you know what is driving the strategic decision-making in the company - it is very hard to automate a decision in an effective way.

DMR: Boardrooms and operations are often at odds in this regard. How do you bring some clarity to the situation?

JT: One of the advantages you get from focusing on operational decision-making is strategic alignment, which always strikes people as odd. They tell us, "We're just talking about high-volume operational decisions here; how does that help with strategic alignment?" The answer is, if you want to change strategic direction, you need to change the micro-decisions that contribute to that direction in a consistent way. The context of an overall strategy is applied not just in executive committee meetings, but in day-to-day interactions on the Web site and in the call center.

DMR: Why is it so difficult to translate strategic goals into operational decisions?

JT: What I see are overinstrumented businesses that have all sorts of infrastructure to understand what is going on, yet no real ability to change the way operational decisions are effected. I'll use a dashboard analogy. I used to work with a guy in the auto industry who'd point out that the dashboard is the flat thing on top that everyone's talking about. The dashboard isn't a steering wheel or a button - you know where you're going, how fast you're going and how much gas you've got left, but you can't do very much about it.

DMR: Are you saying that all the technology gives enterprises a false sense of competency?

JT: I'm saying the people who understand the strategy can't actually go change it because it's not implemented in a way that's successful for them. You have to take all the things that constrain your strategy, plus what your data tells you and find a way to inject it into operational systems. The way a company acts is increasingly driven by its information systems, Web site, billing system or call center system. People click through dashboards, drill down and see a problem. They pick up the phone, call their IT department and ask them to put a change through a system. They call the Web team to get them to do something else, and they call the guy in the call center and say, "I need you to tell everybody in the call center that we're going to do this instead of that." It is just a very squishy process for turning an insight into any kind of effective action. It is interesting that many companies we've worked with for years, like credit card issuers, were able to change their risk strategy and offer credit long before they had reporting infrastructure. It's not like this BI/data warehouse/analytics infrastructure is necessary to do these things, but clearly once you have the infrastructure, the desire to find a way to apply it goes up. People used to say, "We're drowning in data and starved for information." And now it is, "Well, I have the information, but I still can't do anything about it."

DMR: Will service-oriented architectures make it easier to effect changes in decision-making?

JT: The adoption of service-oriented architectures, better CRM systems and better data infrastructures all are enablers of change, but they don't automatically cause it to happen. The basic problem is not recognizing that decision-making is a different kind of automation, and we're still trying to automate decisions using code. If I want a service whose purpose is to decide on a rate I'm going to offer for cross-selling, I have to ask a programmer to change it. I can't tell how that impacts marketing or credit; it is very hard to show whether it's even compliant. If I want a propensity-to-buy model, someone has to hand code it. All these things mean that a service is not inherently more flexible when it comes to changing the way I make decisions. It does give me the ability to reuse the decision and be consistent, but it doesn't let people simply go and change a decision.


DMR: Should there be centralized ownership of all the decision-related things, and at what level would that be managed?

JT: That's a discussion we have quite often. It depends on the class of decision - a credit issuer might have a chief risk officer. Where a class of decision is the driver, the person who owns that context is also often the driver for making sure it is automated in an effective way. As they did with BI, we're starting to see some folks creating decision and rules competency centers.

DMR: Not another competency center! Who's going to pay for that?

JT: (Laughs.) Well, companies have an analytic group focused on predictive models that is separate from the BI group and separate from the groups that adopt business rules. These three aspects need to come together in some way if you're going to deliver real decision automation. The topic of a chief decision officer has come up, and we have a couple of customers who have someone who either has that title or is pretty close. We have a couple of customers with senior people in charge of decision technologies. They tend to be in companies that think of themselves as very decision-centric - credit issuers or insurance companies. They recognize their whole business hinges on a couple of key decisions that they make in very high volumes. These companies are more decision-centric than process-centric. Other industries face a battle to get people to think of decisions separately. You end up with two different approaches. Industries that think of themselves as decision-centric will tend to rapidly adopt a model where someone is in charge of the technologies for decisioning. Others will have decisioning spread out across multiple processes.

DMR: We've been hearing about the interaction between processes that drive decisions and decisions that drive processes. Can you explain that?

JT: At the Gartner BI show, someone spoke about turning processes around so instead of your process definition driving the way a transaction was handled, the data that comes in on the transaction determines what kind of process you do. You invert the process - instead of saying, "Here's my process," you're letting someone say, "Here's my transaction, go ahead and process it."

DMR: What is the benefit of that?

JT: It makes you more decision-centric; you're making decisions all the time about what this transaction wants, says and does. In the other case, the process is still fundamentally in control. The models work in different situations. The mistake people make is they think about data, process and decision-making separately when it is its own genuine skill set and approach.

DMR: Do you think data and analysis people are still pretty separate from the business?

JT: You see the stress and strain between IT departments and the business users. The organizations that focus on decision-making and decision automation as a separate class of problem can look at those disconnects and find a place where they can at least start to solve them. I wouldn't think for a moment that merely thinking about decisions is the simple thing that will magically solve these problems. But it gives you a place to go.

DMR: At the other end of the automation curve, it has a lot to do with the kind of people you hire and what their skill sets are. They have to be multispecialists but they also have to understand the business.

JT: Exactly. You've got to look for technologies and approaches and ways to help bridge those gaps. You've got to find people who aren't just interested in the best fiscal analysis or the fastest code but are thinking about how they can make better retention decisions or make more customers happy. Those are problems people dealt with before technology came into the mix, but it is a particular problem when you're thinking about decision technology because decisioning is core to the way a business runs, very complex and impacted by regulation. You really do need to find a way to bring those people together and give them a common basis for talking about and solving those problems.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access