Pitfalls of process mining tools and ways to deal with them

Process mining tools provide visual, data based and objective insights in to the “as is” process. Typically all the IT systems/applications generate event logs with information like case id, description and timestamp. Process mining tools rely on these system generated “digital footprints” to arrive at the process model close enough to the reality. These tools provide accurate statistics like actual process flows, variations, exceptions, bottlenecks, frequency, and throughput.

While these tools are great aid to the consultants (including process, business and data consultants), there are  few pitfalls.  Due to these pitfalls, sometimes the real “as is” process may not be accurately represented by these tools. One need to be aware of these and take appropriate steps to properly mitigate the risks.

#1: Manual tasks

  • If some steps/tasks in the process are done manually, there won’t be any event logs captured and thus tools can’t discover those manual steps.

#2: No event logs

  • If any system is NOT generating event log, then these tools can’t discover these process/sub-process. While this is less likely scenario, it may occur in few occasions (e.g. legacy system).

#3: Inadequate event log data

  • If event log of any system(s) lack critical information (e.g. case ID, activity), then this is insufficient for the tools to make sense and present meaningful visualization and analytics.

#4:  Event logs with high granularity

  • If event logs are done at high granularity this may hamper and restrict the process discovery to the high level activity only

Process consultants should be aware of these pitfalls and take appropriate mitigation steps. Some suggested steps are as below:


Data identification and data preparation is a vital step in process mining.  Manual tasks to be discovered by consultants during the process workshops and interviews.  Sanity check on systems in scope can be done to ensure event logs are generated and are of good quality for the process mining tools to make sense. In case some systems DO NOT generate event logs (less likely scenario), explore whether event logs can be created from any other sources (e.g. existing system database or history tables etc.). In the worst case, it’s better to explore avenues for fixing of the system at the earliest. Event logs are important for audit and compliance purpose as well and process mining can be a trigger point to get this issue fixed sooner than later. In case of insufficient or high granular event logs, one may explore the possibility of fixing the insufficient data through various means including discussion with business owners, identify alternate data sources and devise custom queries to pull relevant data from multiple data sources.


Ubiquitous use of process mining in RPA

Process mining tools are effective in discovering the processes and it usage during discovery phase is well known. However, process mining can be leveraged in variety of ways throughout the RPA lifecycle. Here are some potential uses of process mining tools during RPA lifecycle:


A. Discovery Phase

1.Visualize “as is” process

2. Note and analyze vital statistics of the process(like frequency, throughput, bottleneck, variations)

3. Leverage statistics to decide whether to optimize the process or to automate or do both

B. Automation Phase

4. After developing the BOTS, you may visualize the “to be” process and see what you designed is what you developed

5. Effective visual testing tool

6. Analyzing the vital process parameters

C. Support & Maintenance

7. Automatically compare old process model with the new process model and quickly find out where RPA worked well and where it didn’t. Remedy as appropriate.

8. Health check: Regularly(say quarterly) monitor the new process by observing it’s vital statistics and take remedial action

D. Audit

9. Compare reference process model with the existing model

10. Identify process non-conformance(NC) cases


Harness process mining in RPA implementation

RPA is well proven for automating standard, repeatable and manually performed processes. RPA gives many benefits including cost savings, quality improvement, cycle time reduction, better compliance and to better cope with capacity variability (Peaks and Valleys).

The “as is” process discovery phase is traditionally been done manually by consultants. Consultants would conduct interviews and workshops with business people/SME, go through the existing process model and documentation, understand and identify process candidates for automation.

Process mining tools automate the process discovery and progressively gaining traction. Research and Markets projects CAGR of 50.3% growth for process analytics market for next 5 years. Process mining leverages the data science and AI to analyze the system event logs and automatically identify the real process model that’s in use (as opposed to the documented reference process model). These tools present excellent visualization of process maps to transparently and easily understand complex process flows easily. Various analysis including throughput, bottleneck, conformance analysis, replay, animation etc. could be quickly done using the process mining tools. Some of the tools help with generating the process documentations as well.

There are various process mining tools available in the market including Celonis, Disco, Minit.io, Epiplex, QPR, ProcessGold and ProM (For academic purpose) which is worthwhile to explore. Interestingly most of the process mining companies have relation with Eindhoven University of Technology (Tu/e) in one form or the other as process mining research primarily emanated from Tu/e.

Lately leading RPA vendors have realized the benefit of process mining. UiPath has recently entered in to strategic partnership with Celonis to offer integrated solution. Kryon offers home grown process discovery as part of its RPA product suite.

Process consultants can leverage these process mining tools and techniques for the process discovery phase to bring tool based and data driven analysis approach in identifying right processes. This way, you are automating the crucial phase of RPA! Isn’t RPA all about eliminating manual process and automating stuff!!

Kill the Golden Goose?

IT service/BPO provides who have entered in to long term contracts before the advent and maturity of RPA, may have low YOY productivity improvement (3% to 5%) factored in to the contract.

It will bode well to implement RPA to improve overall productivity and cut costs. This would demonstrate proactiveness, show genuine desire to transform and delight the customer despite revenue cannibalization. The probability of winning the extension is far higher this way than doing half-hearted attempts or worse reactively proposing transformation options during the rebid process. So, don’t kill the golden goose.

RPA Myths

Are you thinking RPA implementation as pure IT related only? Are you projecting the FTE reduction as the only RPA benefit for your business case? Think again.

RPA should be part of your organization’s holistic digital transformation strategy helping in process optimization, automation, quality improvement, cycle time reduction and enhancing customer satisfaction. This requires involvement from various stakeholders including CxO, Business, IT, HR, Compliance and Security.

Sunday Salon Saga

• It’s a once in a month ritual for me to take my two son’s out for hair dressing. There are couple of decent salons in my immediate neighborhood which I have been a regular. For whatever reasons my son’s don’t prefer them and are very specific to go to a chain salon that’s couple of kilometers away from my home. I have tried to persuade and influence them to go to one of the neighborhood salons as they are at walkable distance, no appointments required and guys there do a decent job to my liking. However, my sons have been insisting on the little far away chain salon and I came to know the reasons later for their liking as good ambiance, no waiting time (i.e as per appointment), cleanliness and the hair dresser is more talkative and kids-friendly.
• While my elder’s son was at the chair and undergoing his hairdressing, I noticed a gentleman in the next chair who appeared to be more interested in his conversation with the hair dresser(let’s call him Appu) than the hair dressing itself. He was very chatty and both of them were discussing about the family, his recent Bombay trip and the just released cinemas. Appu seems to know what the gentleman exactly wanted including the make of the hair-dye, how the mustache should be done etc. I learned that the gentleman Is from Anna Nagar and he has travelled 12kms to this Mylapore salon so that he can have his hair dressing done by Appu.
• I realize the job done by Appu was nothing extra-ordinary but the entire experience he offered to the customer is wholesome and he gained full customer trust. Hair dressing is a commodity business and each street in Indian cities has couple of salons. So, How do you differentiate in a commodity business?. Below are Key lessons from Appu:

  • Know your customer
    Appu knew his customer requirements and likings very well. He seemed to genuinely care for the customer, went beyond the business and established chord with the customer
  • Engage with your customer and offer overall customer experience
     It’s not just about the hair cut but about the overall experience and customer intimacy. Customer is willing to pay more for a good experience for a commodity service

I was telling my son’s that the Rs 50 tip that Appu got from the gentleman was an over-tip but my son’s disagreed. So what’s your take on this?

Quick Tips On Improving CSAT

Customer Satisfaction is the one of the most vital parameters that every project manager tracks and strives to keep it in good health. However, as a project manager how do you keep the pulse of the customer?. How do you know what customer likes and dislikes?. One way is to depend on the Project level CSAT and the Account level CSAT scores and feedback. However, these are lagging indicators and sometimes in may be too late to act. So, how do you keep the customer pulse on a dynamic and proactive basis? Below is couple of suggestions:
• Stay connected to the ground – Stay grounded and ensure your team interaction levels are high and internal review rhythms are followed. More often you will hear the issues internally and by promptly addressing it you can avoid customer escalations.
• Rely on multiple channels and informal approach – Don’t just rely on formal meetings to get all the feedback. Often the informal chats and indirect stakeholders can give you the revealing insights that you can leverage on.
• Mine Escalations – Quality and quantity of the escalations are an important parameter that every PM should have an eye on. Though this is a lagging indicator, look for patterns in escalations and address the root cause. E.g. Do you see a pattern in the escalations that’s related to basic delivery process? Then you need to strengthen your processes.
• Gauge Interaction/engagement levels – Interaction levels and engagement depth are key factors that determine customer satisfaction and interest levels. Are you invited for important meetings or proposals?. Is your opinion respected and counts?. Are you informed or consulted on important decisions? Gauge where you are in the customer decision making cycle and try to place yourself in the upstream.
• Leverage 4 quadrant framework – I suggest to use the below 4 quadrant framework to keep track of the positive (frills, delight) and negative parameters (annoy, disgust) on a monthly basis. The below diagram depicts CSAT in 4 quadrant framework for airline industry which can be easily adopted for you project needs as well. Actively work on eliminating or minimizing the annoying and disgusting factors and maximizing the frill and delight factors.

CSAT 4 quadrant

Blog Stats

  • 8,717 hits
© Dhandapani Ammasai, 2011. Unauthorized use and/or duplication of this material without express and written permission from this blog’s author is strictly prohibited. Excerpts and links may be used, provided that full and clear credit is given with appropriate and specific direction to the original content.