1. Peter Schooff
  2. BPM Discussions
  3. Tuesday, August 26 2014, 09:44 AM
  4.  Subscribe via email
Do you think process mining should have a higher profile in both business and BPM?
Tim Bryce Accepted Answer
I assume you are talking about documenting your existing systems. If so, Yes, it is necessary to identify the processes as well as the data used. By doing so, it is relatively easy to identify the strengths and weaknesses in your systems, and what to improve. Also, by cataloging the processes and data, you go a long way to avoiding redundancies thereby saving time.
  1. more than a month ago
  2. BPM Discussions
  3. # 1
Peter Johnston Accepted Answer
I liken it to the change from discrete, individually wired components on a circuit board to a silicon chip.

BPM was designed in the days of the circuit board. A few process steps, each with a discrete data source, like a battery wired to the circuit board. Not round-trip either – it didn’t really use the data it generated to optimise itself, learn or accommodate capacity.

The biggest benefit of BPM over older methods of process improvement is the instant data the software provides. That data should be working like a car ECU, constantly optimising (in a car it is thousands of times a second) the flow. And it should, like the car, be taking data direct from the source and amalgamating it, comparing it with past performance etc. to create an intelligent, 3D map of what is happening and continually improve the resolution to build a 4K 3D image of what is going on, how it is improving or decaying and where further improvements can be made.

By turning our processes into this 3D datamap of the enterprise we completely obliterate the old, hard wired, dumb programmes like ERP and CRM, which only feed back to us what we put into them without adding real intelligence along the way. We create a compelling argument for the company to base itself round the data as the single version of the truth - the ever flowing, adjusting and improving map of enterprise performance.

So of course we need to include process mining and every way we can of improving our data. Without it we may as well be back in the days of the soldering iron, creating the non-joined up unintelligent enterprise.
Dynamic Process
Oxfordshire, UK
+44 (0) 1491 874368
+44 (0) 7590 677232
[email protected]
Too bad that ooutside the process illusion people just don't work this way. And process mining has no means to automatically provide a 3D-map. How would it do that? What data do you want to improve? It can't even tell you which information goes together and make up a process. It has no means to identify why people do something and what the outcome of what they did was. So why bother ...
  1. Max J. Pucher
  2. 3 years ago
  1. more than a month ago
  2. BPM Discussions
  3. # 2
Max J. Pucher Accepted Answer
Blog Writer
There is no should or should-not as it would all depend on the achievabe benefits.
I have made my position on process mining known some time ago in this post:

Process mining tries to solve the complexity of analysing processes but it starts in the wrong place. Process mining assumes that doing business defaults to rigid processes and that is the fallacy. You can't mine what is not there. Process mining tries to solve a BPM methodology problem and not a business problem. It fails miserably. People need an environment that makes their work easier and simpler and more productive and does not try to turn them into fools with tools.

Process mining offers similar illusions as the Big Data hype. You can't mine what isn't there. Big Data has one use only and that is to track the individual without regard for privacy. It can't actually be used for much else. All other analysis provides useless statistical correlations that do not improve the quality or accuracy of predictions. Process mining has one use only and that is to identify repeated process interactions. It is useless to do so, because when you have already solid repeated electronic interaction then there is no need to turn this into a rigid flow. It already works. The claim that the process becomes easier to change is a lie. It is much easier to send my email to someone else than to ask a process bureaucracy to make a change to my process. Process mining does further not provide any knowledge about WHY something is done and what the goals are. The effort to do process mining across disjointed data sources adds on to the effort and cost to define a process and makes it even less profitable. Don't waste your time and money.

What makes any large information setup valuable is not what is mined from disjointed, uncorrelated data, but purely the information that people provide into a homogenuous infrastructure that they all use. It does not matter if it is Facebook, LinkedIN, or the Waze app where user information identifies traffic problems. Process management would benefit from the same. But it is not 'Social-BPM' because that still requires experts to define the processes at the end, just as with process mining.

Therefore one needs a homogneuous process infrastructure that business people can use without having to define abstract processes, but where they can perform the work and thus they provide all the information about it. Most of it they provide anyway. Some information as to what the goals and outcomes of a process is would be additional. Email and MS-Office can't be used because they lack the ability to describe process information and business data and content in a standard way. In such an environment one can create reusable process templates without any mining. At the same time one can use pattern-matching functionality that will identify user actions in relationship to certain process patterns and events. Eventually it will make recommendations which actions are repeatedly used. If users reject such recommendations then the pattern match is improved. All IT has to provide is the infrastructure and the data interfaces to silos and archives. Business experts can improve the templates that users created while doing the work. The pattern matching engine will continuously learn from user actions and there is no need for knowledge engineers and complex data collection and processing.

All this can be substantially simplified if the process environment allows the business to define the terminology used to perform the process.This is stored in an ontology and allows the system to automatically distinguish between business data and process information.

PS: I am not having wet dreams about what could be but what I propose is available as standard products. Process mining is NO standard product but a complex project that requires many manyears of expertise.
  1. more than a month ago
  2. BPM Discussions
  3. # 3
Emiel Kelly Accepted Answer
As being an employee of a process mining tool supplier, I've been playing around with this stuff for 6 years now. I've always been very skeptical about it and still am.

For sure, it has a high cool factor. 'Wow I didn't know we were able to execute spaghetti with our systems' or 'See those cool colored balls flowing through my process map' .

But it has it's limitations as we are only looking at the system side of a process. And as we all know, process is more than systems.

You can also having a hard time getting useful log files as real processes probably are supported by more systems. So in practice, most time is spent on preparing instead of the real mining experience.

Having said that. In comparison to 'traditional process mapping', the nice thing is that you can focus on process performance (throughput time, waiting time etc). That's more valuable than watching process pictures, but it's but it's no magic because it doesn't tell you why 60 % of your cases weren't on time.

It might give you an indication, but for sure it will not tell you the solution.

And is it useful on day to day basis enterprise wide? Not sure. But if it is, it turned into good old process monitoring, doesn't it?
Sharing my adventures in Process World via Procesje.nl
  1. more than a month ago
  2. BPM Discussions
  3. # 4
Michael Lee Accepted Answer
Once a process is in place and all of the proper data is being logged it would seem that process mining could provide a better ability to implement a continuous improvement project. When mining the data makes visible a high rate of exception handling in a process, for example, then the process can be altered to better handle those exceptions making it more efficient and productive.
  1. more than a month ago
  2. BPM Discussions
  3. # 5
Keith Swenson Accepted Answer
Process discovery during a BPM initiative suffers from Process Confabulation. We may be experts in what we do, but when your expertise involves tacit knowledge, we can't always explain how we do it. If you are getting paid a decent salary, you can't admit that you don't (explicitly) know your job, so you make stuff up. There is no intent to deceive -- it is a natural result of our storytelling brain. You will explain what you truly believe you and others around you do --- only it does not always match reality.

Some people are self-critical, and think they are inadequate and that they should be doing a better job. These people will tend to describe the better job that they think they should be doing. They don't want to admit that they are cutting corners.

When you have someone who is well aware of what they do, they often present the sunny day example of the process when all goes well, because they assume that the reason they got into problems in the past was a failure to implement the process, and an automated process will avoid those problems, and therefor we don't need to include any special considerations for the difficult cases.

If you do happen to try to ask people about the exceptional or failure cases, memory tends to draw up the most familiar cases, and does not give you a reasonable representation of the real distribution or frequency of such problems.

For all of these reasons, Process Mining (a.k.a. Automated Process Discovery) can be a great tool for uncovering what really has happening. Your fact-based, egoless, discovered process becomes the starting point, that everyone can cooperate around improving.
  1. http://social-biz.org/2008/05/25/process-confabulation/
  2. http://www.fujitsu.com/global/products/software/middleware/application-infrastructure/interstage/solutions/bpmgt/bpm-services/apd/
  1. more than a month ago
  2. BPM Discussions
  3. # 6
Jim Sinur Accepted Answer
Blog Writer
I believe that process mining is necessary for a reality based view of operations, but it gets a bad rap as being too complicated and difficult. It doesn't have to be difficult and the results are pretty impressive.

Business operations are based on the collective activities of your people, systems and constituents and improving those actions can prove to be very profitable. However, there are problems with knowing what really goes on without precise and detailed measurements. These measurements are often time consuming and require specialized skills, if you can get at the measures. Traditionally business professionals have to rely on process or technology professionals who have methods that are less than perfect because they rely on interview driven process models to lay the baseline for improvements. While there may be precise measures for individual parts of a process, getting a complete end to end and precisely measured process is usually not possible without close and automated inspection.

Processes are a reflection of the business operations in an organization. By basing process improvement on real metrics or measurement s and allowing business professionals to direct the creation of the existing base line for improvement, the results should be profitable and timely. Improvements driven by these accurate pictures of actual behavior can be designed and implemented in the new and improved processes. How this is accomplished makes all the difference in the size and speed of the benefits. Operational Intelligence can be implemented in several ways, but it is always based on real data.

There are any number of vendors that provide this kind of easier process discovery, one of which is giving away free white papers on the topic listed below. I have also included one of the many case studies I have seen on this topic.
  1. http://jimsinur.blogspot.com/2014/04/ptiney-bowes-canada-delivers-better.html
  2. http://www.stereologic.com/
  1. more than a month ago
  2. BPM Discussions
  3. # 7
Eric A. Stephens Accepted Answer
Blog Writer
I'm interpreting "process mining" to be a baseline state (EA) or define/measure (DMAIC) activity in order to build some sort of improvement program. I (and others too) typically advocate for a "just enough, just in time" approach to understand the current state before venturing into defining one or more target states. Relying only on an automating mining of digital information (SIGINT) may not tell the entire picture which necessitates a HUMINT - human-based context - to be introduced into the conversation. One approach is for the practitioner to define a hypothesis based on "SIGINT" and then vet with stakeholders to ensure a proper course of action.

And yes, I promise to stop reading so many Tom Clancy novels :D
Eric A. Stephens | Oracle Enterprise Architect Director
Oracle North America Technology & Government Consulting
IT Strategy & Architecture Services
  1. more than a month ago
  2. BPM Discussions
  3. # 8
David Chassels Accepted Answer
The core to our research that started over 20 years ago was about how people actually work in any "process" So in a way was “mining” to find out how process actually worked to deliver outcomes. The summary in this research paper published last year gives flavour of what we found http://www.igi-global.com/chapter/object-model-development-engineering/78620 Yes less than 13 generic work tasks can address all business logic where a "performer" (people and machines) create all information Call this stage 1 as the fundamentals of understanding “Process Mining”

So stage 2 in "Process Mining" is applying to build any identified process big or small. This requires step by step all the actions and links to data etc to be identified and custom configuration of the prebuilt generic tasks as the required business logic all linked exactly as required; all in business language.

Stage 3 could be “Mining” a Process to “dig deep into” the process activity to track any problems to allow transparency in resolving operational issues

Stage 4 could be the “Process Mining” constant review of performance of the working process and allowing rapid change driven by users.

So to answer the question yes as I have described very important to business addressing the BPM principles. BUT I remain to be convinced “Process Mining” is “tag” that really adds value or knowledge for the business user indeed might add confusion...as clearly I am!
  1. more than a month ago
  2. BPM Discussions
  3. # 9
Alberto Manuel Accepted Answer
Blog Writer
In my humble opinion, as a practitioner with deep expertise in the field, far beyond discovering process models, I tend to think the possibilities are almost infinite, that goes from information system management to complex social network analysis and knowledge discovery.

For reasons, I assume, related with the low maturity of the approach, much is being discussed about the importance of the most used scenarios, process discovery and compliance checking. Despite this is important for companies that struggle to understand reality and what must be done in order to support change and transformation, process mining can drive innovation how operations are executed. Still, one of the challenges is process mining vendors keep aligning the offer with the classic approaches evolved by Professor Wil van der Aalst, when it should pursue a combination strategy with other discovery and visualization approaches, because not everything, not every operation is process. Something new is coming to the market, but unfortunately is for academic demonstration purposes only.

I see in the future process mining being able, if the vendors, but mostly, the managers (yes the humans) want to embrace the concept of real time enterprise. This is not a fable, is a reality in industry sectors like utilities, water and distribution and oil & gas, but some challenges still exist regarding how to embrace unstructured data and dynamic behavior. To end, I would like to see process mining embracing approaches like the ones used on morphogenesis.

I leave here to pointers:

Fiat Lux – the rise of the real time enterprise

Process Mining must embrace complex adaptive systems
  1. more than a month ago
  2. BPM Discussions
  3. # 10
  • Page :
  • 1

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