Start a Discussion
BPM
user-pic

What is the value of process mining?

Vote 0 Votes
There have been several interesting blogs on the recent Process Mining Manifesto (PMM), from Max J. Pucher to Keith Swenson, but first and foremost, what is the value of process mining?

9 Replies

| Add a Reply
  • In short, I think that process mining as a principal concept will be extremely important in the future but rather than following the ideas of data mining, it has to provide a real-time capability that assesses people interactions to provide guidance towards well-defined business goals and gathers contextual information from the interaction-content to support decision making.

    As long as process mining is used to try and find well defined flow-diagrams in people interactions it just makes BPM more time and cost intensive. Process mining in an existing BPM setup is not really mining but monitoring.

    I have no idea what the value of process mining should be if all it does is to try and automate people interactions by means of flowcharts. The only benefit would be cost reduction like orthodox BPM, which is negative management action to bolster stock market relevant profits.

    The future must be to help businesses to be more innovative or to align strategy and execution and both needs people empowerment and not process automation.

    I propose all these great minds in process mining should put their heads to that target.

  • Hi Peter,

    Thanks for your interest in Process mining! I believe that because process mining is a generic technology (just like data mining) it must be put in a concrete context to highlight its value. The specific benefits that process mining provides vary depending on whether you use it, for example, for increasing operational efficiency, for risk management and assurance, for reducing errors, or for controlling partners.

    In my view, process invisibility is a major driver for process mining (see also our blog post on that topic here: http://fluxicon.com/s/5e).

    I would be happy to chat more about it.

    Best,
    Anne

  • I have had direct experience in using process mining with real customers and real data. Whatever I say about it here is going to sound like hyperbole -- too incredible to believe -- however the results are nothing short of amazing. Organizations that have been struggling for months or years to understand the cause of their inefficiency, have quickly (a few days of work) uncovered facts about where the process is going wrong, along with hard numbers showing precisely how much waste there is.

    The value of process mining is simple: it gives you an accurate, quantitative picture of what your organization has been doing. That picture almost always includes obvious areas of needed improvement. Knowledge is power.

    This is not a futuristic technology, but very useful today.

    The latest version of Fujitsu's product announced this week combines process mining with continuous ongoing monitoring and alerting:

    http://blog.interstagebpm.com/2012/02/07/taking-process-intelligence-to-the-next-level-interstage-bpm-analytics-v12/

    Everyone always asks for ongoing monitoring, but quite frankly I don't think this is where the big value is. You can only monitor what you can anticipate. The big value I have seen is when the picture of your organization includes something entirely unexpected. During mining, we will see something that causes an organizational member to run from the room saying "I can't believe that" only to come back later after verifying with other sources saying "It is true, we really do that." The value is not that it produces a process diagram automatically. The real value is the insight that is gained from a having a detailed picture.

    Think of it as a kind of organizational MRI. A doctor can guess at the problem using external symptoms, but an MRI can give them a picture of the internal situation. It is not perfect, and it does not automatically prescribe a treatment, but just having an accurate picture is sometimes all you lack.

    I participate with the IEEE Task Force on Process Mining and the widespread opinion in that group is that to most of the public, process mining just seems too good to be true.

  • I've worked with some process mining tools. Sometimes it looks like magic, but I think using process mining for process discovery is a start but not enough, because the mined log files might not be about a complete process. And you will miss parts of the process that are not executed by the system (some tools allow you to complete your process model manaually)

    It might be useful for analysis (costs, risks etc), but the most value for me is using it as a monitoring tool.

    Of course then you need a worklow or casemanagement tool to execute your process. If you mine this system continuously you will see your running cases mapped on a process model.

    To me this process view combined with live information is a valuable tool for managing the process.

    You will see workloads, which paths are the busiest. Of course you must be able to act on this information, but at least we are looking at a process then.

  • And now I see, Keith doesn't agree with me ;-)

  • Whatever our initial motivation for automation of business processes, the "surprise" upside is the data that then becomes accessible. Before long ad hoc searches of the data are commonplace and executive decisions increasingly dependent upon them. Indeed, fast, flexible and no-coding-required reporting is becoming a top criterium in tool selection.

    Process mining shifts the emphasis from the data content but to the data flow. By examining which paths data actually follows, what volume of data flows through a given node and by observing the preponderance of choices made in the decision trees it is possible to extract the "real" process and optimize. The "surprise" upside when we process mine is the ability to predict where data will originate and what path it will follow thus enabling us to further optimize and load balance our automations.

    But the key to all of this is that it must be a real time view of our processes with the ability to "see" into the future. It must raise alerts and notifications when processes are predicted to reach predefined thresholds.

    When selecting a BPM/BPA solution be certain to add this to your criteria because, as we become more and more dependent on our process automations, we will need to become more adept and optimizing them real time.

  • And Max also ;-)

    But I am a believer that only execution matters. Every tool, method, etc must be used with execution in mind.

    Staring to those nice flashy mining output? Why? I think executers, process managers must be enabled to continuously adapt their process to deliver the results.

    Ans mining might be useful for that. But to me more in a monitoring way than a "look back what we did wrong way".

    More entrepreneurial than analytical I think. Maybe that's why I sell cars ;-)

  • user-pic

    Process mining is invaluable but can be costly, i.e. analysis paralysis, if an organization is not careful. It is so interesting to me that BPM projects start with the notion that great savings can be achieved by automating a process and the process to start with is one that everybody knows. Six sigma, lean, flowcharting etc. are approaches we use to bring more visibility to what we know. These tools tell us how we can refine the process and make things efficient within the respective process. How many times is that process quantified in the greater context of the organization's business and evaluated relative to the other many processes ? It is only evaluated against what is known. Or for that matter, once that process is automated in some fashion what should be the next process to automate ? Do we really understand all the aspects of the process ? Bring in to the picture social media and social behavior and it gets even more interesting in how to define the process. How many organizations have asked what don't we know, and can we gain insight into those hidden processes and the value to the organization. Sure, there are plenty of disciplines that help define the right 'known' process to automate but once an organization is thru their early stages of the BPM journey, it gets harder to get the big lift and to 'quantifiably' identify where to further enable the business within BPM technology. Process mining is a great way into hidden processes, identifying patterns of value, quantify (as Keith said) value and validate where to prioritize scarce resources. The challenge is finding the right methods, tools and determining at what point diseconomies of scale creep into the mining activities. We have only scratched the surface.

  • Pity that Process Mining is being targeted as a process discovery technique because this reduces the full spectrum of process mining capabilities that goes far beyond that perspective.

    As a practitioner and with hands on experience on process mining project in complex environments like Airport Management (check the case study I carried together with Fluxicon on BP Trends)

    http://www.bptrends.com/publicationfiles/BPTrends%20Article%20Process%20Mining%20-%20Ana%20Aeroportos%20de%20Portugal-Alberto%20Manuel%20%20V01-Final.pdf

    shows quite clearly the value process mining is capable of.

    Thus instead of making reflections about the manifesto, people should start thinking that as results are being delivered that process mining is not a theoretical approach that is trying to become hype, is currently an approach to business transformation and adaptation that goes far beyond traditional approaches available so far.

    I should say that for my point of view, process mining is fact based (not assumption based) combined with human reasoning and business knowledge allow to explore the roots of process execution hidden on today's enterprises due to the fact that information does no longer resides within the company boundaries, and also because business dynamics is making impossible for us humans understand the myriad of execution complexity on a standard basis (to many systems, to many exceptions, to much information).

    Process mining delivers a palette of approaches that goes far beyond structured work. It's possible to analyze dynamic processes. The fact that these methods aren't property or do not sit on particular process mining tools, but are available to practitioners is driving to many people crazy, but this is not a weakness, actually is the beauty of process mining, open, tech agnostic, delivering results quickly where it was impossible before to do.

    Regarding the fear as expressed Stuart, the never ending story project, that is not a problem of process mining, that is a project management issue regarding scope and effort management.

    We are not scratched the surface we are already shaping the future.

Add a Reply

Recently Commented On

Monthly Archives

Blogs

ADVERTISEMENT