I’ll admit that I see things through the narrow perspective of my kind of work and experience.
So when I say that I have yet to see a real need for BPM simulation, nor a successful implementation of BPM simulation – I might be missing the bigger picture.
If I was a process manager in a very large organisation with only a few processes to look after and plenty of time on my hands – Sure, I’d love to have a simulation tool, all of my own, something to play with, something to keep me busy, something that makes me look valuable to the organisation.
But if I’m a process manager in a company that has just started implementing BPM, then simulation looks like a bit of an overkill. I’d be looking to implement as many processes as possible, get visibility into what is happening in my organisation.
Spend time on implementing new processes more than building what-if scenarios.
I’ve been asking myself these questions:
• Does simulation add a lot of value to the process or just tweak it?
• Is simulation much better than runtime trail-and-error?
• Is simulation a great time saver or money saver?
• Would the organisation be saving great amounts of time and money, or just a bit.
I think the most important article on BPM simulation was Rashid Khan’s The Hype about Simulation and Optimization which discusses the pro and cons of the concept.
There is a common belief is that the market is not yet mature enough for simulation.
I think it’s mistaken. The truth is that BPM simulation is not yet mature enough for the market. It’s missing the last mile.
Think Star Trek. Captain Kirk asks the computer a question and gets an answer. Why can’t process owners do the same?
Why do we let simulation gurus, smart as they may be, touch the system. Why add human error to the equation?
Let the simulator engine generate a list of recommendations to solve the business pains, based on process data from the database:
Where do resources need to be added to shorten process time, which business rule has the highest failure percentage and needs to be looked into…
Is BPM simulation really needed or is it just a “nice to have” feature?
Why can’t it be automated?
Adam Deane 
Yes Adam, we need simulation, even primitive without complex performance calculations. For example, I found that “running” tokens through a model is the best way to better understanding of the model. See http://www.slideshare.net/samarin/practical-process-pattern (download before viewing to see animation).
Thanks,
AS
By: A.Samarin on 17/11/2010
at 12:13 pm
Adam, my experience with a lot of BPM implementations over the years is that simulation is used much more often in the vendor demos than in actual practice. Although I’ve seen some mature complex processes benefit from simulation for fine-tuning performance, in general I don’t see it used much during initial process design.
By: Sandy Kemsley on 17/11/2010
at 1:09 pm
In all my years in workflow and BPM I’ve never had much faith in simulation – it usually demos well but beyond that I don’t think it’s much use – here’s what I think
http://pykesplace.blogspot.com/2009/11/who-needs-simulation-anyway.html
By: Jon Pyke on 17/11/2010
at 3:39 pm
Adam, thanks for a real world analysis. Simulation is even worse than the original design. It creates the illusion that the processes will work this way. It takes the ultimate step and replaces users with simulated ones. And that should somehow help?
Simulation clearly shows the main purpose of orthodox BPM – automation for the sake of cost cutting. I can’t even think of a single process that would benefit from simulation. Alexander, who is an unwavering believer in rigid processes, said the right thing: sometimes one would even understand the process without simulation. I suggest you get immedatialy RID of that process and create one that is understandable by people.
Rashid Khan proposes that you have to make 150 assumptions for a fairly basic process to simulate it. Lets face it: a most probably higher number of assumptions was made when designing the process!!! But lets say there are 300 assumptions that can be a right or wrong combination of guesses in any perceivable way. The outcome? UTTER NONSENSE!
It makes a lot more sense to use a process environment that enables the workers to perform processes the way they are needed and not the way they were designed and simulated.
By: Max J. Pucher on 17/11/2010
at 9:09 pm