In Part 1 of this blog series, we introduced the topic of Process Mining and how it will deliver the return on investment that SAP clients have been pursuing for 30 years, how process mining works and some of the Software tools to enable Process Mining.
In this blog (part 2), we will look in depth at Celonis and how to apply Process Mining to your SAP system.
Why Celonis?
There are over 20 Process Mining tools on the market. While each of the tools has a different niche, as an "old" SAP consultant, I have found that Celonis is most closely aligned to analysing SAP processes in an efficient manner.
Therefore, I would use this to tool to apply process mining in an SAP context. They have prebuilt analyses for the major SAP business processes (e.g., P2P, O2C etc) and have standard extractors for SAP source systems (e.g., ECC, S4HANA, SRM, CRM). Additionally, they have added out-of-the-box content for common business issues in each process. (e.g., in P2P they have 3 way match, late deliveries etc).
Can’t I do this with existing SAP Toolkits?
Over the years, SAP has provided various approaches for process analysis such as BI content, Spend Analytics and Operational Intelligence. While each of these provides value, these tools do not have the specific Process Mining capabilities to “Discover the real processes”. This is the breakthrough approach!!
What outcomes can I achieve using Celonis Process Mining?
The range of possible outcomes that can be delivered by Celonis Process Mining is as broad as the number of processes within your organisation! In this blog, we will focus on the following 3 areas which have proven ROI for existing clients;
how to identify non-compliant processes at first glance,
how to optimize throughput times by identifying bottlenecks and
how to reduce manual interventions in process execution through automation.
To illustrate these benefits, lets walk through an example based on SAP P2P data.
The Analysis – Variant Explorer
Once we have uploaded the activity log and constructed the data model in Celonis we are ready to start analysing. In the following examples we have used the “Variant Explorer” to understand the process.
Variant 1
In Variant 1 we can see the happy path. This is the most common path through the process.
The numbers on the activities represent the number of cases and the numbers on the connections represent the average throughput time
This “Happy path” represents 1 of 655 variants which covers 38% of the cases
Variant 2
Adding Variant 2, we see an alternative process which starts directly with the creation of the Purchase Order. While this is a common approach, best practice is to use the Purchase Requisition as the starting point for the process.
The first 2 variants represent 2 of 655 variants which covers 51% of the cases
Variant 3
Adding Variant 3, we can see the activity “Change Price”. This is a common example of rework. Eliminating or at least minimising “Change Price” represents an opportunity to streamline the process by reducing the throughput time and also the cost.
The first 3 variants represent 3 of 655 variants which covers 64% of the cases
Variant 4
Adding Variant 4, we can see a common non-conformance. In this case we see 74,632 cases where the starting point is to scan the vendor invoice. For most organisations this is in breach of purchasing policy and may signal a potential fraudulent situation.
In this example we have switched the metric to show the number of cases flowing through each path.
The first 4 variants represent 4 of 655 variants which covers 71% of the cases
As we can see from this analysis, the typical P2P system will have a large number of process variants.
The fact that each activity has a timestamp means we can generate performance metrics and highlight the activity connections with the highest throughput times.
This represents a massive opportunity for clients to streamline their process and remove bottlenecks!
How does my Actual Process compare to Best Practice Models (Compliance / Conformance) ?
Many organisations have their theoretical best practice models developed in static tools such as ARIS, VISIO or PowerPoint.
These can be loaded to Celonis and compared against the actual data to show a level of conformance
In this example we can see that our actual data “complies” with the Best Practice Model 57% of the time and there are 15 process “violations”
Violations
A violation is either an activity which is not expected (e.g., change price) or an activity which is out of sequence (e.g., starting the process by scanning the invoice)
The top 3 violations are “Change Price”, “Scan Invoice” and “Block Purchase Order”
From here we can drill down to further analyse the top violation “Change Price” which represents 14% of the cases
Automation
Additional KPI’s are available to show the impact of this non-conformance on throughput time and the automation rate
In this case we can see the violations are causing the throughput time to increase from 29.0 to 31.1 days and the Automation rate to decrease from 50% to 45%
The automation rate can be derived from the type of user performing the activity
Activities which are performed manually can be identified and targeted for automation
So now what? Identification of Root Causes or What to do next?
At this point we have identified “Change Price” as a violation to our standard process.
Based on Data Mining algorithms, Celonis will suggest which data points have had the biggest influence on the violations.
This can be based on the number of cases or the strength of correlation
In this example, we see that Vendor Umbrella Corporation has had 50,716 price changes
Drill down to Individual Cases
The intelligent filtering capability allows us to jump directly from the root cause analysis to the individual cases causing the violation. In this case it’s the list of purchase orders for “Vendor “Umbrella Corporation” where the activity “Change Price” has occurred
At this point we are in a position to discuss with our procurement department or the vendor why so many “change price” activities are occurring
Conclusion
From this discussion we can see that a tool such as Celonis can be used to identify non-compliant processes at first glance, optimize throughput times by identifying bottlenecks and reducing manual interventions in process execution. In fact, we almost have a guided procedure as it assists us to identify possible improvement targets, provides a to do list of process violations as well as a hit list of likely root causes!!
So, that is the end of this 2-part blog on Process Mining. I hope you have enjoyed the blogs and I hope I have stimulated your interest in Process Mining. Stay tuned for further blogs taking a detailed look at Celonis capabilities. Each blog will focus on a business process (e.g., P2P), a common challenge (e.g., 3 way match) or a Celonis capability (e.g., Multi Event logs). If there is a process or capability you would like me to focus on, please let me know and I’ll do a blog on that!!
If you would like to understand what is happening in your processes in a matter of weeks not months then please contact us today at info@TEC150.com and take advantage of our real world experience and state of the art certification.
Best Regards
Comments