Many analysts have at least some level of theoretical knowledge on how to properly apply use case modeling.  Unfortunately, key elements are often missed resulting in a use case model which isn’t effective.

Use case models include actors and use cases.  An effective model must identify both comprehensively.  Yet, analysts usually fail to identify all of the relevant actors. 

To solve this problem, I like to recommend an approach that maintains a separate use case diagram and actor diagram.  Then, the actors from the actor diagram can incrementally be added to the use case diagram as the use case diagram evolves.

Starting with a separate actor model places a focus on actor analysis.  This is the process of identify as many distinct roles as possible within the scope of the business area or system.  Then, each of the actors can be organized by showing actor generalization/specialization (inheritance).  Showing actor generalization/specialization in a separate diagram can help keep the use case diagram clean, since generalized actors may not need to be shown initiating use cases.  However, it can still be helpful to show this hierarchical organization of actors separately as it aids the audience of the diagram in understanding how each actor is similar or different.

The actor diagram is just one piece of the overall actor analysis.  Each actor should be accompanied by a description of the role which can be conveniently maintained in a spreadsheet.  Understanding what the actor does at a high level (2-3 sentences) as well as what the actor does not do (differentiation of actors), helps the audience of the model gain a well rounded understanding of the actors involved.

While the actor analysis is being completed, the use case diagram is also created.  Each high-level use case should be identified.  These are the use cases which directly deliver value to the actors.  Each actor which initiates the use case should be show with a relationship indicating the initiation.  Similarly, other actors which provide support to the use case or receive something from the use case should be shown as secondary actors (non-initiating actors).

After the high level use cases are identified, they can be decomposed to shown smaller portions of the use case which are re-usable across use cases, or that may be used by only a subset of actors.  This is done by using the “includes” relationship between the use cases.

Finally, each use case is incomplete if the analyst hasn’t taken the time to write a one paragraph description of what the use case does.  These descriptions are critical to ensure that the use case model is effective and useful.

Use case descriptions should state:

1. What triggers the use case
2. What happens within the use case
3. When the use case ends
   

By placing a distinct focus on both actors and use cases and providing a brief description of each the effectiveness of the use case model is maintained.