BPM.com
  1. Peter Schooff
  2. Sherlock Holmes
  3. BPM Discussions
  4. Tuesday, 16 May 2017
  5.  Subscribe via email
With the speed of business today, is process simulation still an important aspect of BPM?
Accepted Answer Pending Moderation
As the first run to show users what their input looks like and actually does yes, it is a good start. This then moves on with feed back to the deployable solution and so user buy-in can change attitudes and remove fear of seeking future change.
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 1
Accepted Answer Pending Moderation
There were times when I would have paid a lot for a simulation feature on the process engine.

Now we have an automated testing framework, much more useful for developing solid process-driven apps. And much, much faster.

Would be cool for customers to show them simulations, but it's not critical anymore.
CEO, Co-founder, Profluo
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 2
Accepted Answer Pending Moderation
Simulation - demos well, but of little practical value. Simulations are little more than a stack of assumptions put on top of each other. As such it is easy to prove just about anything with minor changes in some of those assumptions to get to the model answer that you think you want.

True discrete event simulations require a lot of validation work (i.e. effort) to ensure that the stochastic models used actually reflect reality. This is normally a lot more than most business people have in mind.

The only real valid use of simulation models in BPM is to validate or test and assumption, and postulate what might happen. Proving the financial return (of a change) ... is a flawed exercise as the outcome the modeler is trying to show becomes the dominant factor (i.e. they lead to deterministic models).
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 3
Accepted Answer Pending Moderation
Yes, process simulation is still an important part of BPM, for two reasons;

1) quick/easy to compile/roll-out a mapped process and assemble a small group of stakeholders to piano-play your process. If a process has 10 roles and you can only find testers to take on 4 of these, widen the role for the other 6 and have the person who is hosting the session take on those roles.

2) no amount of staring at a paper process map can do as good a job of identifying bad logic, bad routings, and bad forms, as a group of in-line testers.

During early testing, you don't need to design forms and you don't need to automate branching decision boxes.

Images of forms are fine so long as there is one field where comments can be recorded.

Manual branching decision boxes with, again, one field where you can record notes on what will become decision rules are fine.

It's important to be able to be able to re-map\re-compile\rollout processes during testing sessions - you lose a lot of momentum if you have to go away and come back the next day with changes.
References
  1. http://www.kwkeirstead.wordpress.com
Comment
"have the person who is hosting the session take on those roles"
actually, that's what our testing framework does very well - it impersonates those roles through the interfaces (logs in with user credentials, performs tasks according to the test scripts and inputs test data as planned) and performs multiple use cases, testing at the same time the process logic and the interaction to interfaces.

People with more imagination than myself would call it robotic process automation for testing purposes. ;-)
  1. more than a month ago
  2. BPM Discussions
  3. # 4
Accepted Answer Pending Moderation
I think the question should be slightly different – what are “business cases” for simulation capability in BPM?

Let us look at different phases of BPM:

[model/plan phase]
- better understanding of processes by business people (piano-play from @Karl)
- iterative development together with business people

[automation phase]
- testing

[execution phase]
- permanent process instance health check, e.g. are we still in the SLA limits?

[control phase]
- benchmarking of a process instance against good and bad examples

[measure phase]
- anticipation of resource limitations

[optimisation phase]
- comparison of optimisation variants

Thanks,
AS
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 5
Accepted Answer Pending Moderation
Simulation has its place as one of the analytical tools in the BPM toolkit but it is valuable in some limited use cases.. What appears to be more important is the ability to do some form of high level activity based costing using static data.
Comment
Simulation should be the first cut of the eventually deployable application as such not just an analytical tool but part of the "agile" build ...without coding and using business language/logic. Activity based costing follows with real data and any required improvements readily supported.
  1. more than a month ago
  2. BPM Discussions
  3. # 6
Accepted Answer Pending Moderation
Thus spake the Oracle* of Oceanside:

I quibble with the premise, but setting that aside for a moment:

With the exception of certain you only get one shot at this so don't mess it up type of processes (think satellite launch): No.

Automate. Measure. Tweak. Repeat. Start focusing on CX/UX and spend less time worrying about subtle, Tayloresque “improvements” that save you time but cost you customers.

(*) Oracle is a registered trademark of the Ancients of Delphi, S.A. All rights enforced by Cerberus.
http://www.bplogix.com/images/icon-x-medium.png
-Scott
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 7
Accepted Answer Pending Moderation
I have seen users putting a lot of emphasis on simulation and predictive analysis capabilities in a BPMS some 3 to 5 years back. In most cases where these users then acquired such a platform or complementing tools to an existing BPMS, simulations or statistical analytics very seldom took place, though.
Despite that pattern, I firmly believe that process simulations are still very important or better yet, have to become an important part of BPM in practice. In fact, I think that process simulations represent one of the still under-explored pockets in the world of BPM. In order to gain momentum, eventually BI and statistics have to be brought closer to BPM (ex. simple and multiple regressions on process data series and cross sections to project and compare different scenarios as well as process designs).
NSI Soluciones - ABPMP PTY
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 8
Accepted Answer Pending Moderation
In 2004 'yes.' Now, 'no.'
Comment
Hmm. Modern predictive analytics is a type of simulation.
  1. more than a month ago
  2. BPM Discussions
  3. # 9
Accepted Answer Pending Moderation
"Still" or "ever"? Nice demo, but rarely used in practice. Consider also levels of simulation (individual process all the way up to entire business unit). And what is that simulation level that supports business decisions for a given role (the only business case for simulation is the value of better decision making -- but what is that second-order intangible)? There's often too little slack in business manager time to afford the O/H necessary to maintain a simulation programme, except in very defined circumstances. And in those circumstances, you buy Lanner etc. Interestingly, process mining is a step towards simulation. So what about the future? The future must include more business simulation; this is both an opportunity and a challenge for vendors and management alike. Better business decisioning (and making sensible use of all that big data out there) demands simulation. But who will be responsible for building simulation models and fragments? It's an open question. Nevertheless, there's a bright future for business simulation, especially in organizations that want to be successful. Tough sell though.
Comment
  1. John Morris
  2. 1 year ago
  3. #3931
Business simulation is based of course on assumptions. Narrow distribution probabilities are reasonable assumptions in well-defined circumstances which may be found inside corporate boundaries. However, once you get to larger-grained simulations that depend on events outside the corporate boundary, the value of simulation as a contribution to better decision-making becomes increasingly questionable. (We are now in the fat-tailed world described by Nassim Taleb.) Perhaps the challenge of selling simulation is in part because managers intuitively understand the limitations of big-picture simulation. Small-picture simulation can be shown to be be worthwhile though.
  1. more than a month ago
  2. BPM Discussions
  3. # 10
Accepted Answer Pending Moderation
What's the goal of simulation? I think to understand how your process design will work out in the future. Often based on a process model. And that's the problem I think; a process model is not, like many other models, a small or virtual version of reality.

Like the model of a car. Which is really a small version and where aerodynamics can be simulated in a wind tunnel. It will give you a good indication of future reality.

Or flight simulators which are the virtual version of the real reality.

I think that's a little harder with processes and their models. First of all because real life processes have a lot of dynamics which are hard to express in models. Like I wrote here

For sure, some simulation tools can cope with these dynamics, but it's most of the time not so easy to set it up. Getting the data is of course easier these days because of things like process mining. The data that was mined can be used to predict the future. Maybe.

Personally I like more a culture of experiment and improve. And just make real life a simulation.
Sharing my adventures in Process World via Procesje.nl
Comment
RE "And just make real life a simulation." - your poor customers... Glad to know that modern planes and cars are built with a lot of simulation.
  1. Emiel Kelly
  2. 1 year ago
  3. #3933
Alexander ;-) Thanks for enforcing my point as I said that simulation of cars and planes make sense. Because they can be simulated as as small or virtual version of reality.

And those are products, not processes. Models of processes are just some agreements on the meaning of boxes and arrows.

Emiel, OK, let us consider another example - football - which is about coordination of players on the pitch. It seems that teams should not waste money extra training? It will be a serious innovation.
  1. Emiel Kelly
  2. 1 year ago
  3. #3939
Also rigt, Alexander. Those are all good appliances of simulation/Training. I thought we wer talking about simulating some modelled processes by adding some data and then hit the magic button.
  1. more than a month ago
  2. BPM Discussions
  3. # 11
Accepted Answer Pending Moderation
As Alexander points out, the scope of "simulations" is wide.

I suspect stakeholders value "piano play" simulations and management easily understands the benefits of this type of simulation for process design/process mapping.

Except in highly automated processes where most of the steps are performed by user "System", the benefits of advanced simulation (ability to anticipate performance) are often elusive because of certain realities:

a) variability of task performance times

Reason: users typically work on multiple cases and even with a focus on just a few of these, they get called away, get distracted, & get diverted. Because of "S" curve, each time a user suspends work on a step, there is an exit time and then, when they re-visit the step, there is a re-entry time.


b) variability of wait times between tasks that is not easily discernible in data mining/analysis

Reason: Users get alerts when steps are held up/taking too long but they are not always able to immediately act on alerts, Supervisors take time to react to escalation of alerts.


c) insertion of ad hoc steps (e.g. a customer calls, a piece of equipment fails).
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 12
Max Young
Blog Writer
Accepted Answer Pending Moderation
I think that @John Morris has it right when he says "The future must include more business simulation".
Net/Net, if you believe that the visual representation of the business process is the key to what we do, then you *have* to accept that simulation can illuminate the path forward. Heck, I spoke about exactly this just a few weeks ago @ BPMNext.
References
  1. https://vimeo.com/203389031
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 13
Accepted Answer Pending Moderation
BPM.com keeps its finger on the zeitgeist. Coincidentally with this forum topic, today Digital McKinsey published an article in Medium on almost the same subject:

Solving complex planning problems using old math: Optimization applied to operations planning

Note that the word "simulation" does not appear in the article, although the word "process" does (4X). Nevertheless the tools and business needs discussed overlap significantly with the tools and business needs of simulation.

Especially interesting are the author's notes concerning the "why" of the article -- because adoption of things like linear programming and predictive analytics are not nearly as widespread as they can be. (Big opportunity therefore for consulting . . . and tool sellers.)

All this is more evidence for @Max's belief in a strong future for simulation and predictive modeling.
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 14
Accepted Answer Pending Moderation
The answer to this question depends on whether you sell a process simulation tool or not!

My thoughts:

  • There's value in understanding your model. Simulation is one way to understand how it behaves (better). Simulation based on real world data (replaying it and then projecting forward) always seemed more interesting than pure simulation with assumptions.
  • There's value in the technologies that underlay things like simulation and process mining. We've found alternate uses for them that inform what we do.
  • We're not in the business of selling simulation or process mining, per se. We're in the business of solving business problems, and we don't see a lot of customers asking for these things. Selection bias abounds and after having written this i will probably find just such a customer tomorrow :)
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 15
Accepted Answer Pending Moderation
Process simulation is an essential productivity feature for any BPM software. The ability to watch a process running and quickly edit it before its deployment to production significantly simplifies and accelerates development of high quality processes.

Automatic testing of processes is equally important, but it cannot replace quality visualization in simulation. Simulation can be considered as process debugging capability by analogy with regular programming. Of course, it is possible to debug a program or script entirely based on execution of its unit tests. However, it is very handy and helpful when you can visualize running code step by step to quickly highlight problems.

It is also important to have fully featured "step into" capability to stop process on every step and evaluate all its variables and conditions. It implies much more detailed insight than just a basic animation of process flows.
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 16
Accepted Answer Pending Moderation
Yes,From my point of view!
Process modeling is thus an important part for the successful implementation and BPM is an evolution of BPR (business process re-engineering) which dealt with radical changes. One ofthe approaches to evaluate a proposal is simulation.
Comment
  1. more than a month ago
  2. BPM Discussions
  3. # 17
  • Page :
  • 1


There are no replies made for this post yet.
However, you are not allowed to reply to this post.