At the start of every project it is necessary to define the goals of the project and frame how the model will be built to satisfy these goals.  One goal that is implied in every project but rarely discussed is communication.  For a simulation to have any value, the results must be understood and appreciated by the decision makers.  As simulation modelers it is our job to determine the most effective way to get our point across.  The extreme example is GPSS programmers.  As late as the 1980s I would see the results from a model formatted as a 200 page printout off a VAX mainframe.  The customer would need to be able to pour through the data with the programmer and understand the nuances of the model to have any appreciation for the results.

Now that is not going to happen today.  Almost every simulation is either graphical or has the benefit of a PowerPoint presentation to ease the burden of understanding.  However, we still have a wide gamut of methods for presenting results.  In some circles, detailed presentation of confidence intervals, type II error, chi-squared analysis is expected as part of result presentation.  To leave them out will only cause more question and doubt on the part of the customer that the modeler knows what he/she is doing.  More commonly, customers want to know is results are statisically relevant (significant) and they are more than happy to leave the verification to the analyst.

The point here is that you need to know your customer.  Are they the type that want to see the gory details of the statistics?  Can you present terms like statisitical significance without losing them in the process?  Do they want a highly visual model to involve plant floor personnel in the model validation?

We can make simulation models do so many things and present them in so many ways, it is necessary to make the method of communication known from the start.  A few tips to head off problems….

1) Always have detailed statistics available during your presentation in case you are called on to present them.  You don’t need to make them part of your base presentation but they should be available.

2) View the results from the perspective of the customer.  For those of you familiar with Lean Manufacturing and Value Stream Mapping.  Stick with the value added statistics.  Try to tie each slide, graph, or statistic back to the original goals of the model.  Presenting utilization details for non relevant operations just serves to hide the real point you are trying to get across.

3) Don’t feel like you need to overwhelm people with numbers.  In the end, simulations make decisions.  Hit the high points early in your presentation.  Tell them what the results are early and then back it up with the detail.

4) Make things visually recognizable.  You do not need to be explaining what each graphic on your screen represents.  If it is not obvious to someone who is familiar with the system, then you did not do a good job presenting the animation in your model.  Legends, Color Coding, Dynamically changing labels, and visual detail all get people past “what is that?” to “why is that?”

5) It is hard to over communicate in these situations.  Feel free to stress important points several times in several ways.  Just avoid Powerpoint poisoning.  We have all been in presentations with 50 slides that could have been done better in 15.  Ask yourself, why do I have that slide?  If it does not tie back to the original goals, keep it as back up material for those that want more detail.

Good luck and know your customer.