Decision Analysis (Part 1)
What Kind of Decisions?

by Ronald G. Ross

This column is excerpted from Decision Analysis Using Decision Tables and Business Rules by Ronald G.  Ross (Oct, 2010), an in-depth white paper available free on:  http://www.brsolutions.com/b_decision.php

We define decision analysis as identifying and analyzing key decisions in day-to-day business operations and capturing the decision logic used to support them.  Decision analysis should be a central focus of business analysis.  It is a highly effective, business-oriented means to capture certain kinds of business rules.

A decision is a determination requiring know-how or guidance; the resolving of a question by reasoning. 

An outcome is an answer to such a question; that is, some result from making a decision.

A decision task is a task or action in which some decision is made.

The end-product of decision analysis is decision logic in the form of decision structures, decision tables, and business rule statements.  This decision logic is rendered in a form that is practicable (ready for deployment whether to staff or ultimately to machines), enterprise-robust, and business-friendly.

Kinds of Decision Appropriate for Decision Analysis

Decisions appropriate for decision analysis share four essential characteristics in common, collectively called 'DOOR' for short.  Such decisions are:

• Deterministic, rather than intuitive or ad hoc.
• Operational, rather than tactical or strategic.
• Objective (encodable as explicit rules), rather than subjective.
• Repetitive, rather than one-off or infrequent.

Examples:

• Should a customer be considered gold, silver, or bronze level?
• Should an insurance claim be accepted, rejected, or examined for fraud?
• Which resource should be assigned to a task?

Typical kinds or patterns of DOOR-type decisions include classification, evaluation, selection, approval, assessment, assignment, allocation, diagnosis, and prediction.  As these kinds or patterns suggest, DOOR-type decisions always involve some significant determination at a particular point in a business process.  Such determination should exhibit both the following additional characteristics:

• It seeks to identify the best or most appropriate of alternative outcomes.
  
  
• It occurs at a given point in time in a process.

Also, decision analysis always takes an organizational view of decisions, not an individual view.  How an individual makes personal decisions in his or her own life (even at work) is not a concern.  Decisions of interest for decision analysis are strictly those of the organization.  If such decisions are made by individuals, the individuals must be acting in a legitimate capacity and making the decisions on the organization's behalf.

Kinds of Decision Not Appropriate for Decision Analysis

Decision analysis is not appropriate for decisions that are:

• Simply gatekeepers for whether some action can be taken.  For example:  Under what conditions are we allowed to take this order?  Under what conditions are we allowed to charge this kind of expense?  Under what conditions are we allowed to waive this requirement?  Decisions appropriate for decision analysis are about resolving business questions by reasoning, not directly about constraining or judging the correctness of business activity.  See below for additional discussion.
  
  
• Removed from day-to-day business activity or abstract.  For example: Are we following our rules correctly?  Is our business process optimized or well-coordinated?  Are we aligned with our business strategy?  Decisions appropriate for decision analysis are always directly in-line to a business process.
  
  
• Ones that can be resolved or predicted by a formula or calculation — i.e., by some equation(s).^[1]  For example:  How much is owed for an order?  How much business has a customer done in the last 12 months?  What is a student's cumulative grade point average?  Decisions appropriate for decision analysis generally do not have outcomes that can be calculated mathematically, based on circumstances or cumulative history.

Keep in mind that the end-product of decision analysis is decision structures, decision tables, and business rule statements — in other words, explicitly-encoded decision logic that can be understood and managed by business people and business analysts.  Decision analysis might also be suitable where the end-products are statistical models, neural nets, or similar forms of non-verbal representation, but these other forms are not the focus of this discussion.

Kinds of Business Rules Not Addressed by Decision Analysis

Will decision analysis enable you to capture every relevant business rule?  No.  Does every business rule fit into some decision table?  No.  A great number of business rules cannot be captured effectively using decision analysis or decision tables.  Other techniques are still needed for rule capture^[2] and expression (e.g., RuleSpeak^[3]).

Rather than supporting operational business decisions, many business rules are ultimately about business governance.  Such rules do not pertain directly to determining the best or most appropriate outcome among alternatives, nor do they apply only at a single point of determination.  Instead, their purpose is to prevent undesirable situations that could occur at any of various points in time.  In serving that purpose, they act as gatekeepers for business activity at each point in time.  Some simple examples:

Business Rule:  A customer that has ordered a product must have an assigned agent.

This rule is not about selecting the most appropriate agent for a customer, nor does it apply only at a single point in time (e.g., when an order is taken).  Instead, the rule is meant to be enforced continuously — for example, even if the agent assigned to a customer retires or leaves the company.

Business Rule:  A work group posted to a union site must include a union liaison.

This rule is not about selecting the most appropriate work group to post to a site, nor does it apply only at a single point in time (e.g., when a posting occurs).  Instead, the rule is meant to be enforced continuously — for example, if the current union liaison drops out of a work group, or if a non-union site to which a work group is posted subsequently becomes union.

Business Rule:  A student with a failing grade must not be an active member of a sports team.

This rule is not about selecting the most appropriate sports team for a student, nor does it apply only at a single point in time (e.g., when a student joins a team).  Instead, the rule is meant to be enforced continuously — for example, if a student who is already active on some sports team should let his or her grades fall.

Business rules like these generally fit no particular pattern.  For that reason they cannot be effectively managed in a decision table along with other, parallel rules.  Instead, they need to be expressed as individual business rule statements.

Your organization has hundreds or thousands of such one-off business rules.  Decision analysis does not do it all!

Part 2 of this three-part series will introduce business analysts to the basic elements of operational business decisions.

Notes

[1]  Such formula or calculations are often numerous, but generally fit no particular pattern.  For that reason they cannot be effectively managed in a decision table along with other, parallel rules.  Instead, they need to be expressed as individual business rule statements.

[2]  For example:  pattern questions, strategy development, policy analysis, business milestones, and precedence diagrams.  These techniques are supported within Proteus, the BRS methodology for business analysis and business rules.

[3]  RuleSpeak is a set of guidelines for expressing business rules in structured natural language.  RuleSpeak guidelines are available for free in English, Spanish, Dutch, and German on www.RuleSpeak.com.  All statements of business rules in this discussion use RuleSpeak.