Last time, we expanded on what operational decisions are and explained why and how so many of these decisions are hidden.  In this instalment we go on to explore some techniques for finding operational decisions and discover how applying enterprise decision management can deliver smart enough systems.

Finding suitable hidden decisions that have the characteristics of a decision that repays automation and improvement is an important step toward smart enough systems.  Many suitable decisions will occur to you simply because you're looking for them, in much the same way you notice a particular model of car after you own one.  Using techniques such as brainstorming and facilitated sessions can be effective, too.

The most suitable decisions are typically those an organization uses to manage its interactions with associates, although not exclusively.  Here are some straightforward ways to find good candidates for hidden operational decisions:

• Analyze the reports you generate, and find out what prompts action from those who read them.  From this information, you might be able to identify rules for taking the action and have the system use the data to take action for them -- automate the decision that's made when the report is reviewed.
  
  
• Read the procedures or your users' cheat sheets.  If they work around something in the system or are forced to override the system, you might be able to figure out which rules in the system are wrong.  If changing these rules isn't easy, you might be able to externalize them so that you can manage and change them more easily.
  
  
• Processes that involve lots of manual review -- by your auditors, for example -- and are hard to monitor might be worth automating.  Other potential candidates include areas where showing compliance with regulations is a problem.
  
  
• Supervisors in your call center can give you information on what decisions get referred to them.  If some involve no new data collection or are otherwise mechanical, perhaps they can be automated to allow front-line workers to act on behalf of customers without having to refer them to a supervisor.  Any supervisor's decisions might be good candidates for automation.
  
  
• Get the list of actions your CSRs or their supervisors can take on behalf of associates.  Some can be achieved through the IVR system or Web site, but there might be other actions that could or should be.  Often Web sites and IVR systems only collect or report information and leave decision making to people.
  
  
• Analyze users' requests to see what feature they want so that they can self-serve.  Underlying many of these requests are decisions made manually that prevent self-service.
  
  
• Change logs might show you that pieces of the system are always being changed.  The business or decision logic in your legacy applications is often the cause of these high-maintenance components.  You can measure how much time and effort these changes take and assess whether IT is behind schedule in making them.  You could externalize this part of the application and make it possible for the business to change the rules itself to improve this part of the application as and when needed.
  
  
• Check whether your business users get all the data they need the first time they interact with an associate.  If they have to go back and ask for more, or if they can specify what data they want and why and when they want it, you might be able to derive the rules that would let you collect the data they need (but no other data) the first time.
  
  
• Analyze the data you have, and consider data mining or predictive analytic techniques.  Establish what you could predict based on your data.  If this information would be useful in running the business, see whether you could improve a decision being made by using your data in this way.
  
  
• Find out what your business users want to know.  You might be able to find a way to derive this information from the data you have.  You might also be able to find out what they would do if they had this information, and see whether you can automate the action, too.

Like most skills, finding hidden operational decisions gets easier with practice.  It's important to find the operational decisions you're going to focus on before you use an EDM approach to improve them.  To understand this point more clearly, consider a utility company as an example:

• A core function of this utility company is to deliver energy to customers, so it must measure and bill for this energy in a predictable way.  Applying enterprise decision management to this decision -- the amount to bill -- is probably not worth the effort, because it's a mechanical calculation.
  
  
• If the utility company offers credit terms to customers, using enterprise decision management in the credit department for decisions on who should be offered credit, how much, and at what rate is a suitable choice.
  
  
• Even if credit isn't offered, the utility company needs to decide how to treat customers who are behind on their bills -- which ones should have extra time, which ones should be called, and which ones should be referred to an outside collections agency.  This business decision also would be suitable for enterprise decision management.
  
  
• If the company adopts 'smart meters' that allow different pricing for electricity at different times of the day, its pricing decisions and customer segmentation decisions (for billing plans, for example) are good candidates for applying enterprise decision management.  Pricing will go from simple and mechanical (with the old meters) to dynamic and complex.  Working out pricing that incents off-peak usage profitably is hard and results in much more complicated pricing decisions.
  
  
• The utility company's maintenance and repair operation is a constant trade-off of staff, contractors, overtime, and priorities and would contain other candidates for automating decisions.

Other kinds of organizations have other suitable decisions.  Marketing and promotion decisions in almost any organization are good candidates, because often many choices and options have to be considered before the right offer is made.  Product configuration and pricing decisions are also good candidates, especially when the product or pricing model becomes complex.  Routing and shipping decisions in an automated supply chain or logistics environment often involve complex sets of rules and regulations and a lot of relevant information.  Picking targets for audits, benefits eligibility, tax processing, and regulatory enforcement are examples of government decisions.  Most organizations have many suitable decisions buried in their operations, whether or not they are aware of them.

Enterprise Decision Management and Smart Enough Systems

Enterprise decision management means taking control of operational decisions and automating them.  Table 1 shows how this approach can be used to achieve all the characteristics of smart enough systems, as discussed in an earlier instalment.^[1]  Unless your organization's operational decisions are automated intelligently and thoughtfully, you will struggle to make your systems smart enough to meet current and future demands on them.  Enterprise decision management gives you the approach you want to deliver the systems you need.

Table 1.  Enterprise Decision Management Delivers Smart Enough Systems

Enterprise decision management, although an effective approach, isn't suitable for every type of decision.  Taking the characteristics discussed previously and other aspects of decision making, you could summarize the appropriateness of enterprise decision management as good, moderate, or poor, as shown in Table 2.

Table 2.  Appropriateness of Enterprise Decision Management for Types of Decisions

To Summarize:  In this instalment we have explored some techniques for finding operational decisions and discovered how applying enterprise decision management can deliver smart enough systems.  Next time you see when to use and not use enterprise decision management, when we take a look at the ROI for Enterprise Decision Management.

References

[1] James Taylor and Neil Raden, "The Need for Smart Enough Systems (Part 2)," Business Rules Journal, Vol. 8, No. 8 (Aug. 2007), URL:  http://www.BRCommunity.com/a2007/b358.html  

# # #