Model Detail

[MJMcClellan]

We began modeling about eighteen months ago when we began designing "to-be" processes as a part of "greenfield" project. Now as we approach the implementation of the first of several new systems, the question remains what is the appropriate level of detail to model to. We want to use the models as input to development and monitoring (BPEL & KPI's), aids in training, staffing level planning, procedure management, compliance (SOX) and possibly others. It is not yet clear to me what the appropriate view is. Clearly a 40,000 foot view is too high, but how deep should we go?

Thanks,

Mike

[ms1mm0]

Hello Mike

I guess the answer is "it depends". It depends on what you want to achieve from the model. For some people modelling is a "process of discovery".

When you state you are about to implement then I assume you mean deploy to a process engine or run time.

Assuming you have defined what the business benefits and objectives of the business processes are - that will define the level of granularity you need to model to.

For example some customers want to automate parts of process to achieve
* high throughput by streamlining the process
* ensure consistency and regulatory compliance
* ensure that service levels and customer satisfaction is met
* reduce cost
* enforce work management policies and procedures.

Only when events are collated from all the events that occur as part of the process will you be able to judge whether the benefits you were aiming for have been met.

One assumes that your processes are designed to meet some business goal.

So before you deploy - simulate your processes to understand how they will perform in your business environment. This will demonstrate whether or not you have modeled to the right level of granularity. It will also reduce business risk by identfiying the "unexpected" before it happens in your production environment.

Hope this helps

Mark S