Resolved
3 votes

I've seen a few mentions recently of process mining. Do you see process mining playing an ever growing role in BPM going forward?

Tuesday, October 27 2015, 09:46 AM
Share this post:
Responses (13)
  • Accepted Answer

    Tuesday, October 27 2015, 09:53 AM - #Permalink
    Resolved
    3 votes

    Only as an opportunity for the vendors to charge an arm and a leg. Beyond that, like most things "analytics," few will do it, fewer still will do it well.

    • Emiel Kelly
      more than a month ago
      Agree, as a 'standalone' thing it doesn't make much sense. Should be tied to the execution of processes ans show performance indicators that make sensen
    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, October 27 2015, 10:07 AM - #Permalink
    Resolved
    4 votes

    I think Process Mining will play a big role and won't be called process mining as it does. When the products come out that leverage process mining it will be buried behind other names or labels (cognitive computing anyone?) but the hallmarks of process mining will still be there.

    Like
    1
    • E Scott Menter
      more than a month ago
      I'm interested in what hallmarks you're thinking of, if you don't mind commenting.
    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, October 27 2015, 10:12 AM - #Permalink
    Resolved
    4 votes

    It definitely will play a bigger role indeed. However, and that's IMO the crux of the matter, it needs to be positioned as a part of the whole. Currently you have to be pretty careful with positioning process mining as some sort of a holy grail. I mean: it's a snapshot, can be pretty incomplete (danger of sub-optimization). I do look at business processes as not just transaction based only: what to do with the white space?

    I do believe that more sophisticated technology can and will make a difference here. But as long as people (emotion, politics) are involved I wouldn't put all my analysis bets on this interesting mining technology... :-)

    Like
    1
    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, October 27 2015, 10:13 AM - #Permalink
    Resolved
    3 votes

    Did some projects with it in the last 7 years. Although some think it is still new, to me the hype is a little over. Always been very sceptical about it, because it just shows bi in a process oriented way.

    if you don't have the right data, it will not add much value. Besides that its only shows you symptoms of process performance, not the causes.

    because of better availabilty of data, it might get more usefull the next years, but in the end it's just a tool to support process improvement with the common drawback that you improve in the locker room, not during the game.

    Like
    2
    • Walter Bril
      more than a month ago
      Agree with you Emiel, although I do believe that -eventually- (realtime) mining will play a bigger role. I'm sceptical though and perhaps a bit worried about "innocent DMU's" blind following algorithms (read: great sales presentations) which only some uebernerds comprehend :-).
    • Patrick Lujan
      more than a month ago
      It's only as good as you tell it to be and what is done with it.
    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, October 27 2015, 10:22 AM - #Permalink
    Resolved
    4 votes

    Process mining is an exciting, growing science. Its potential is significant, but its true power lies in the ability of the vendors / consultants to actually use process mining as a tool for continous improvement and decision / operational support. I dream of this as Operations Research 2.0.

    I think it can be used during the game as well, I can imagine a simple scenario of load-balancing human resources and other resources during process runtime (referring to your article today, @Emiel)...

    But I'm afraid that a potential hype will motivate clueless IT marketing people into devoiding the concept of any meaning. Just as it happened to IoT, Big Data etc.

     

    Like
    1
    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, October 27 2015, 10:28 AM - #Permalink
    Resolved
    3 votes

    I don’t think process mining will have a role by itself. Maybe it would be a nice capability of big BPM Suites, but not as a standalone discipline or tool. I think that process will be more agile, and measuring and discovering how the run should be too.

    So, if process mining evolves greatly, and can be used without consultancy services, it has a future as part of SaaS or agile BPMS’s.

    But, in a similar way as BI failed to explode its potential without consultancy services, if process mining is not enough easy and out-of-the-box ready to use, it won’t be massively adopted, no matter how you name it.

     

    • Bogdan Nafornita
      more than a month ago
      Juan, I don't know how BI is supposed to help without strong analysis skills and relevant business insights. I'm quoting John Morris here: "Data doesn't analyze itself"

      Hence the fallacy of IoT / Data Mining etc - their implementation requires deep domain knowledge. This is not something easily found outside people's brains.
    • Juan J Moreno
      more than a month ago
      Bogdan, I strongly agree with you, and that's my point: as BI does, process mining requires strong knowledge, or "consultancy services" to be useful. And this is what keep it from being massive. I'm not saying it is nos useful, on the contrary, it is, but I think there's not going to be a market boom of process mining specialized products.
    The reply is currently minimized Show
  • Accepted Answer

    Maria Paz
    Maria Paz
    Offline
    Tuesday, October 27 2015, 10:42 AM - #Permalink
    Resolved
    3 votes

    I think that process mining is a useful tool that hasn't reached its full potential yet.

    Most KPIs used in BPM are useful to identify bottlenecks but fail at discovering how those processes are executed. Process mining, if used right, can identify alternative courses, patterns and common mistakes. This improvement opportunity can be filled with process mining.

    Like
    1
    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, October 27 2015, 10:44 AM - #Permalink
    Resolved
    3 votes

    Process mining is indispensable for understanding customer’s journey (which is also a process – see ref1 and ref2).

    Thanks,
    AS

    The reply is currently minimized Show
  • Accepted Answer

    Ian Gotts
    Ian Gotts
    Offline
    Tuesday, October 27 2015, 10:45 AM - #Permalink
    Resolved
    3 votes

    Process mining is good for telling you what people are doing now, not necessarily what people should be doing now. But with the "digitization" of industries the strongest role for process definition is understanding what people should be doing in the future. Creating that blueprint for the future, in process terms takes business owners + (virtual ) conference room + strategic thinking + whiteboard (or tech equivalent).

    So it this context process mining is "fighting for a better yesterday" whilst the winners are "designing a better future".

    Like
    2
    • Walter Bril
      more than a month ago
      I have been thinking; it might help in faster (cheaper?) AS-IS discovery. But still with all the current worries I already mentioned above...
    • Emiel Kelly
      more than a month ago
      Then I'm still afraid that it's not AS-IS, but Has-Been.
    • Scott Francis
      more than a month ago
      sounds like you guys have *never* worked on a project where someone told you how things work today (the baseline assumption) so that you can design a better tomorrow... only to find out that their description of "how things work today" was completely wrong in really critical ways that would inform a better product direction?
    • Emiel Kelly
      more than a month ago
      Sorry Scott, might be because I'm not native English, but I don't understand what you are trying to say.
    • Walter Bril
      more than a month ago
      Scott: Actually on the contrary :-). The point I'm trying to make here is that AS-IS process discovery / capture (e.g., "how does it work today") can be absolutely supported by technology such as process mining. And exactly to avoid the problem you describe.

      However, process mining (as it currently works) is IMO only helpful when it comes down to (transaction based) systems. When you leave out the non-automated factor, you might / will miss important white space stuff.
    • Emiel Kelly
      more than a month ago
      Have to agree with Walter. With process mining you are very dependent on the log files whether or not you look at real process. Parts might be missing, but also: actions for a case might be done in 6 systems. Then you have to combine it all to find the complete process.

      Still data preparation is the biggest effort for process mining. When systems get more process oriented (like a BPMS), value might be higher.
    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, October 27 2015, 01:09 PM - #Permalink
    Resolved
    2 votes

    Yes where data is buried in the "mine" and is required for businesses operations so the process will automatically trigger the digging out of required data. Process will also control creation of new data as one version of the truth begins to reduce unnecessary duplication and thus storage requirements.

    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, October 27 2015, 04:52 PM - #Permalink
    Resolved
    3 votes

    Nice metaphor from @Emiel, to wit "improvement happens on the field, not in the locker room". I think thought that this metaphor is actually an argument for process mining - because PM is very much about making changes in near real time "on the field".

    There is suspicion reflected in several comments that the past is not a good basis for predicting or modeling the future (which is also a fallacy inherent in a lot of big data hype). This is a valid concern, but I think it misses the power of what PM is all about.

    All current BPM (or even all IT) is at heart "Platonic", with executive gurus imagining a future, top down. Process mining on the other hand as it develops would be expected to be deployed for all managers, including front-line managers. You could even say that process mining is the first "technology of the tacit".

    And there's no reason that PM can't be integrated with modeling (e.g. something like Lanner). Process mining is dependent on logs, but with the rise of IoT and cheap sensors, it's hard to imagine there will be a shortage of data. PM itself is still in early stages; the software at this point can only handle a single dimension of data, when reality of course is multi-dimensional.

    I expect that PM will become a feature of all BPM products, and certainly consultants will dine out on it. But insofar as BPM is at the core of the work of business, over time I expect process mining to become the core of BPM. Process mining is surprisingly easy to use, and the reaction of business people to a PM demo is a joy to see.

    The challenge for champions of process mining technology I think is mostly around business readiness. If BPM puts pressure on people to think systematically about their "deep domain knowledge" (pace @Bogdan), then PM really puts pressure on executives. No more "wait states on the golf course" while IT figures out how to implement your latest idea.

    Problem today? Solution tomorrow. Implementation the following day.

    And again and again in a process mining-driven fast-cycle environment of constant business evolution.

    Enthusiasm aside, it will likely take years for process mining to be absorbed by executive cadres.

    • Emiel Kelly
      more than a month ago
      To me it's still locker room talks as mining will only make sense with log files of finished cases. So. those case have been executed well or not and you can discuss them after the game.

      During the game would be real time monitoring to me. Pretty sure you can use mining technology for that. That is what process management to me: real time insight in the progress of cases, so that you can still act when something is going wrong.
    • John Morris
      more than a month ago
      It's quite possible to run case or process progression visualizations in real time on existing models based on process mining. It's even more exciting if one adds prediction. There are opportunities for management and staff adjustment in real time - which I think puts process mining "on the field".
    • Emiel Kelly
      more than a month ago
      But when process progress visualization is shown on existing models, it is just monitoring, isn't it?
    • John Morris
      more than a month ago
      Emiel, good point, and gets to how process mining and "as-is BPM" shade into each other. The difference is how the process models are constructed -- "by themselves" (i.e. via process mining math- and graph-based software technology) or by massive "human effort". The cost of traditional modeling is so very high, and we all know the results.
    • Emiel Kelly
      more than a month ago
      Yes, I know the marketing stories behind 'Automatic Business process discovery' ;-)

      No serious; when you see mining for the first time (for example on a conference) you might think Wow, a model by a click, performance information on top of it and those flowing balls....I want that!

      But in real life you are very dependent on the data in the log file to :

      - discover a real process (trigger till result)
      - discover the whole process (different systems used, manual steps)

      And besides that, the workflow is not only the process. Some mining tools can also add the executors to the model, but things like dataflows ; I've not seen them.

      Although it is cool technolgy and it might show more process performance information then we've ever seen, it is still symptoms.

      - 30 % of the cases were too late
      - 2% of case followed a strange route
      -Cases waited (on average) 4 days in front of step X

      Cool to know, but just symptoms. More important is to know why and even more important what to do about it.
    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, October 27 2015, 06:55 PM - #Permalink
    Resolved
    2 votes

    I'm not a process mining expert, and I don't even play one on TV. As a result, I won't speak to process mining specifically; however, I agree with the gist of Ian's position above (at least, as I understand it). To wit: everybody is way too focused in general on the "as-is" or current state of a process. Sure, there is some information that's worth knowing—for example, who the stakeholders are, or where problems have generally arisen—but in general, I find the obsession with the precise documentation of a process you're about to replace a little odd.

    "As-is" research is like preparing to drive to work on the first day of your new job by pulling up Google Maps to examine the route to your previous job. No doubt there are some streets in common, but you've been driving those roads for a long time. Surely it would be better to focus on where you're going, rather than where you've been.

    The reply is currently minimized Show
  • Accepted Answer

    Tuesday, November 10 2015, 04:53 PM - #Permalink
    Resolved
    1 votes

    One of the challenges that process mining faces is it did not move to hot or real time analytics, despite it having that built-in capability.

    As it becomes more important to monitor operations in real time, it is critical, from an adoption perspective, that vendors integrate process mining in their monitoring toolset.

    These words I wrote some years ago continue to feel fresh today:

    Organizations live in a world where interdependence, self-organization and emergence are factors for agility, adaptability and flexibility plunged into networks. Software-based information systems go into a service oriented architecture direction and the same goes to Infrastructures where services are become structures available in networks. inspired into empirical studies of networked systems such as Internet, social networks, and biological networks, researchers have in recent years developed a variety of techniques and models to help us structurally understand or predict the behavior of these systems. Those findings are characterized by been supported on the “complex networks” concepts

     

    The reply is currently minimized Show
Your Reply

Join the Discussion

Want to join the discussion?

Login, or create an account to participate in the forum.

Top Participants

Dr Alexander Samarin
277 Replies
28/09/2016
David Chassels
270 Replies
28/09/2016
Emiel Kelly
222 Replies
28/09/2016
Bogdan Nafornita
209 Replies
28/09/2016
E Scott Menter
182 Replies
27/09/2016