Reflections on Business Process Modeling, BPMN, and BPMS
By brian | April 27, 2011 | BPM, Open Source
Last week I attended Bruce Silver’s online BPMN course. In the past, my BPM Software company purchased versions of his online course for some of our lead business process consultants and process architects. This time, though, I decided to take the time to personally attend his online course. The course took place for about 4 hours per day during 3 days and was offered using an online remote learning tool and consisted of lectures, exercises, and questions and answers.
First, let me say that the course was excellent, and anyone working with BPMN today should have this course as a foundation. Besides being a great presenter/lecturer, Bruce is one of the gurus of the industry. You really feel like you are getting the benefit of years of experience direct from someone who has “been there and done that.” The purpose of this blog post, however, is not to give a detailed overview of the course (popular BPM blogger Sandy Kemsley was also in the online course and wrote a very detailed review which you can read here). Rather, the purpose of this post is to provide some perspective on the course from the vantage point of a Business Process Management Suite Vendor.
As Sandy notes, it is not easy to block 4 hours a day for three days straight, and the temptation to multi-task is always nagging at one’s heels. However, there is nothing like getting a little distance from one’s daily labors to gain a new perspective on BPM Software and its unique challenges.
Here are a few of the reflections and questions that I have been thinking about after taking the Bruce Silver course:
1. The Vast Majority of those using BPM Software are not using a BPMS
2. Massive Room to grow? Why not previously?
3. BPMN is too complicated!
4. No BPMS can get around the need for a clearly communicated and defined process
5. Greater need for BPMN Style Training
1. The Vast Majority of those using BPM Software are not using a BPMS
Early in the training, Bruce made a statement to this effect (I apologize if I am not getting the quote exactly correct, but the essence of this statement by Bruce was that BPMS is a fraction of what is going on in the world of BPM). His statement caught me a little off guard. My company’s open source business process management software gets downloaded hundreds and often thousands of times a day from around the world. So, as you can imagine, this tends to go to our heads. We think that everyone who has anything to do with bpm, bpm software, or workflow software wants to implement an automated Business Process Management Suite. After all, who would go through all the trouble to create a “non executable workflow model” when you can create an “executable” one? Who wants to create just a BPMN diagram when it is possible to create an executable version of the same process and then run the process in an automated environment where everything can be managed, monitored, and improved?
Well, apparently the answer is – most people using BPM technology.
Of course it wasn’t difficult to believe Bruce’s statement, but none the less it is cause for reflection. Software companies tend to have a lot of ego, and we believe the world revolves around our vision. It’s just part of our business and we tend to think others share our same philosophy. But Bruce’s comment made me think about modeling in a very different way, and made me realize just how specialized business process management software and business process management suites are even today.
2. Massive Room to Grow? Why not previously?
If most are using BPM Software to create diagrams and not executable models then this begs the question – why? And does this mean there is going to be a tremendous amount of growth in BPM Suites in the immediate future? If so, why did this growth not happen previously?
Certainly the obvious answer has to do with the standard. BPMN 2.0 was only finalized in August of 2010. Issues of execution were never addressed in BPMN 1.2. And as has always been the case in technology, it usually takes open standards for there to be exponential growth. For companies to invest in technology (either time or money), there needs to be some assurance that the investment will be protected. Open standards provide this assurance.
So, yes, with the standard for executable models now in place, we should see a significant uptick in growth…with a few caveats.
3. BPMN is too complicated!
The fact is that BPMN is too complicated. Most of the BPMS Vendors that are implementing BPMN (most are doing this), are engaged in the same race to the bottom as before. BPMN for many BPM Software vendors is simply a long and extremely complicated list of RFI/RFP check boxes. As is usually the case, one vendor sets the RFP and the rest race to be able to check the same requirements’ boxes. The end result is that vendors continue to produce extremely clunky, bloated software that focuses on adding features without purpose and without thinking about the end user.
Bruce made the point many times that there really is only a core set of elements that need to be used in most diagramming situations. In fact, I would go a few steps further and state that only a few core elements should ever be used. Instead of providing more options, BPMS vendors will at some point need to come to terms with how to get users to design processes better and more clearly. More is not usually better – especially in software.
4. No BPMS can get around the need for a clearly communicated and defined business process
We and our partner network work with customers all over the globe to help them automate their processes in our BPM Software. The simple rule is that the software always works but the process does not. In other words, the software is generally the easy part –understanding and then modeling the process is the difficult part. As is often the case, the software vendors want to try and solve everything with software, but at some point the software becomes a distraction. Does a typewriter, a laptop, or an ipad help a novelist produce better novels? No. Today we have got lots of easier and better ways to actually write the novel, but the difficult part, i.e. the writing of a great novel, has not gotten better because of technology.
5. Greater need for Style Training
What has helped more novelists produce better novels is better technique. And, as any college graduate knows, the bible of better technique is Strunk & White’s Elements of Style. Bruce gets this (and has positioned his book to become the Elements of Style of BPMN), and after 12 hours in his course I can understand why he doesn’t seem so interested in the BPM Suite itself. The fact is that the BPM Suite tends to work just fine whether it is ProcessMaker, Pega, IBM, Appian, or another. The problem is much deeper and needs to be solved in a much less automated way. The problem is the “style.” This problem needs to be solved through better education and better teachers. We need to train better process consultants and this is a long and arduous journey with no shortcuts. Ok, well, almost none.
It turns out, if you use ITP Commerce’s modeling tool, it actually does a style check sort of like the grammar check in word, and actually tells you if you are following Bruce’s style guide. This is not a bad start, but I used grammar check in word for my last short story, and the New Yorker still hasn’t accepted my submission. 
