How to Get Started with Process Mining? 10 Key Learnings by Dr. Lars Reinkemeyer


Hello everyone,

I recently found a very interesting post on LinkedIn. Dr. Lars Reinkemeyer, Global Process Mining Lead at Siemens, shared his experiences regarding the question “How to get started with Process Mining?”.
I really enjoyed reading it and most of his observations definitely look familiar to me! What do you think? What are your learnings when it comes to implementing process mining in an organization? It would be nice to start a discussion on this topic as I think it’s crucial for all of us :slight_smile:

"Taking Process Mining to a global level – 10 key learnings

“How to get started with Process Mining” is a question which I am regularly asked: having built a global community in a large global organization - with several thousand colleagues working regularly on a couple of hundreds data models - has allowed me a couple of interesting observations, which I would like to share:

  1. It all starts with … scepticism: coming up with a new topic like Process Mining in a large and complex organization more often raises concerns than excitement. I.e. Managers who have been in charge of Business Processes for a while tend to assume that their processes work as designed and tend to be sceptic against any transparency deep dives.

  2. It’s all about people: the best tool will fail if it is not picked up by the right leaders in business. Finding the right change agents who understand how process mining can e.g. improve operational efficiency or drive Digitalization is crucial. These colleagues are the real process heroes, who will take the tool to build something innovative and make the difference for the company.

  3. Visualizing complex processes with floating dots … is a nice teaser: showing actual process complexity, nicely visualized with floating dots always causes high Attention and Interest. Moving on to the Desire and Action is the challenge, which can only be conquered with the right people and right use cases.

  4. Go for sprints, but prepare for a marathon: sprinting for a first minimum viable product is important to keep customers excited. Our first sprint in 2014 took 14 days and produced a small demo to show delivery orders cross-unit and cross-system. Since then this project has developed into a de-facto standard reporting with a large number of users and budget.

  5. Good is good enough: don’t go for the last bit of perfection but rather focus on the big picture and the big levers. My observation was, that some people tend to search for single inconsistencies within millions of events, while others look at the big picture e.g. to drive digitalization – with much faster and higher impact.

  6. One size does not fit all: interestingly many functional departments have chosen different approaches, even if processes such as customer orders and procurement orders appear to have many similarities. While e.g. Order Management chose to go for global standardized KPIs, procurement went for a consultative approach with individual focus on every unit.

  7. Provide an open Platform and build a strong community: complementary to building corporate standard reports, we have invested time and effort to build a strong community. This includes e.g. training of >200 process mining analysts, regular info sessions and social media communications. As a result, amazing use cases come up regularly from different parts of our organization.

  8. Process Mining is great, combined with Reporting it becomes even better: pure process mining, i.e. working with a process explorer only, will not generate full impact. A KPI reporting for management, combined with a process explorer will enable the organization to get crucial insights and derive operational actions.

  9. All you can eat fosters appetite: having bought a corporate license very early in our journey was a strong argument to go for maximum usage of our platform and making it a preferable solutions compared to other BI Frontend tools.

  10. RPA, Predictive and AI are buzzwords, but do not (yet) rock the organization: along with the public excitement we have built reports e.g. to predict late deliveries, monitor process conformance or use order automation rates as a heat map for applying RPA. Each of these showcases has generated high management attention, the operational impact remains to be harvested."

(Dr. Lars Reinkemeyer, 2018, LinkedIn)