Decisions Add Value To Business Rules

Ron Ross recently published a great post Business Rule Means These Three Things. In it he said it is critical that a business rule is a rule, is under business jurisdiction and is practicable. In the comments, Ravzan Radulian asked about these things in the context of decisions and DMN (the Decision Model and Notation standard).This is a great question so here goes:

The first key element of a rule is that it is a rule - it acts as a "guide for conduct or action". Decisions involve the selection of a course of action from a range of possible actions. Decisions are often guided by rules - rules defined by policies, by regulations or by best practices. The relationship here is clear - the rules guide or direct or constrain how the decision should be made. In DMN decisions are decomposed and modeled in a decision requirements model and then the logic for each decision in this model is expressed using a set of business rules. Decisions in DMN go beyond rules in two important ways, however:

• Decisions can be made in other ways, especially in this era of analytic and cognitive computing. A decision might be made analytically. For instance a decision might be to decide how likely a customer is to cancel their subscription in the next 30 days and this would be determined by analyzing data not by specifying business rules.
• Because decisions actually select an action they are more useful for injecting intelligence into business processes. The choice of action made by the decision steers and controls the process. BPMN and DMN have a clear and unambiguous relationship - a task in BPMN invokes a decision in DMN. How the decision is made, what sub-decisions must be made first and how business rules are applied is all handled by the DMN model, simplifying the process model.

The second key element of a business rule is business jurisdiction - the rule must be owned, defined and managed by the business. While one can use declarative logic to manage technical rules, when we talk about business rules we should be talking about business rules! Decisions similarly should be business-focused. While decision modeling and DMN can be used to model technical decisions (data transformation etc), its true focus is on business decision making. Business owners find DMN decision models easy to read, easy to build and easy to discuss. DMN decision models can be built quickly and iteratively to build business understanding. Decisions in DMN have some advantages over rules in this regard.

• Decisions and their outcomes are always directly tangible to the business and broad sections of the business will have strong opinions about how decisions are made. In contrast business rules are often lower level and too fine grained to be widely-recognized by organizations or their clients.
• Decision models provide an iterative, top-down approach to add detail. Each layer of detail in a DMN decision model makes things clearer and each "leg" can be worked on separately. Business rules can hide the forest behind the trees and trying to find all the business rules in an area can be very waterfall-y.
• Finally the structure of a DMN decision model provides a framework to manage business rules capture. A DMN decision model structures business rules in a decision-centric way rather than an organization or policy-centric way.

The final element Ron focuses on is that rules must be practicable. Decisions must likewise be practicable. Decisions in a DMN decision model provide a framework also for assessing the practicality of a rule and sets of rules. A set of rules is complete when it accurately and completely defines how one decision in the model is to be made. Rules can be individually practicable but it is their ability to define a decision that makes them collectively practicable. Linking the rules to the decisions they implement and those decisions into a logic, declarative structure using DMN is critical. Two further notes on decisions:

• The buck stops with a business decision: it has to yield an outcome in very case. An individual business rule, on the other hand, may have not cover every circumstance or may raise exceptions. A set of rules can be defined to make a decision and that set becomes the unit of value.
• Decisions have a clearly defined business value, often expressed in terms of having an impact on a KPI or a goal. How well it meets this goal needs to be monitored (and improved). This gives them an extra degree of practicality.

If you are interested in how decision modeling really works and how business rules fit into thise context, Jan Purchase and I have just finished our book, Real-World Decision Modeling with DMN (sign up to hear when it hits the press). He and I did a video on the differences between decisions and business rules and you might also enjoy this post on 3 reasons rules architects adopting decision modeling.