System transformation
with process mining

The problems of a system transformation in detail

A central problem of a system transformation is the lack of knowledge about the exact process flow. If you are lucky, there is already a documented actual process. However, processes usually do not run as documented. There are variations and exceptions that have arisen due to evolved structures and inadequacies in the existing system. If you try to record all process variants using traditional process interviews, you will quickly reach your limits. This is because this knowledge is often spread over many heads and requires a huge number of interviews and process modeling in order to get close to reality. And in the worst case, there is no defined actual process as a starting point. This then has to be reconstructed from the current way of working. This makes it difficult to determine the requirements for the new target process.

This in turn causes the operational risks mentioned at the beginning. This is because you run the risk of overlooking important requirements that are critical for smooth operations. This can lead to interruptions or failures in business operations after the system has been introduced. These in turn are associated with lost sales, customer dissatisfaction and reputational damage. See also the Handelsblatt.

Scarce resources are also a factor that should not be underestimated. On the one hand, traditional process mapping requires employees with an understanding of the entire end-to-end processes. They also need the specialist knowledge to map them methodically and correctly, for example in BPMN notation. On the other hand, employees from the business areas and IT also have to spend a lot of time explaining the processes and being available for interviews and queries.

However, the clever use of process mining can eliminate many of these problems or minimize the necessary use of resources. This will help you reach your goal faster and more efficiently.

Supporting data-driven transformation throughout the entire project with process mining

If you first need a general overview of the topic of process mining, I refer you to our article and our white paper State-of-the-Art Process Optimization with Process Mining. There we explain what process mining is in general and what possible applications there are in the various areas of operations.

A note in advance: Process mining is of course only as good as the quality of your data and the consistency in your systems. For end-to-end mapping of processes, all relevant steps must also be available in the system. This does not even have to be in the same system. When changing systems, however, it is necessary to have a clear assignment. But even if you are still a little weak in this area, the use of process mining can be worthwhile. Sub-processes are examined more closely and the data gaps that are discovered can be addressed directly during the transformation.

So how exactly can you use process mining to support and safeguard the transformation project? We can initially divide the transformation project into three phases: the preparation, the actual system implementation and the adaptation and change phase after the system implementation. You can achieve benefits in all three phases by using process mining. And the best thing is that everything builds on each other so that no work has to be repeated or redone. As the project progresses, you need to enter less and less input into the system and can generate more and more output.

Preparation phase: Pre-migration

Process Mining provides you with a visualization of all process sequences contained in the system, allowing you to analyze and evaluate all process variations. Many individual variants are coincidences or the result of human error. However, in the case of frequently occurring variants that deviate from the current actual process, it is worth taking a closer look at the causes. Find out why these variants bypass the system. The causes are often new business requirements that have arisen over time but cannot be mapped in the old system. This is where you will find your first opportunities for improvement for the future target process. This also gives you the opportunity to quantify the potential for improvement in terms of process time saved or errors avoided. This allows you to define initial requirements for the new target process.

Fit-gap analysis and benchmarking with process mining

When introducing a new system, it is often attempted to work with standard or best practice processes from the system provider. The implementation costs are significantly lower here, as there is no need for time-consuming programming of individual functions. With process mining, you can benchmark these processes against your existing processes and carry out a fit-gap analysis. Where do the standard processes meet your requirements and can they be adopted without major effort? Where is a slight adaptation of the standard processes necessary and where is the gap too large to work with standard processes? In these cases, you can subsequently evaluate whether custom programming or adapting the workflows to the standard processes is the more economical solution. This is also a good method for comparing different system providers. In this way, you can find the one that offers most of the required functions out of the box.

If the system is being introduced across multiple locations or affects several business units, the preliminary analysis also offers the opportunity to carry out internal benchmarking. Different process flows for the same task can be easily identified. This allows you to strive for harmonization even before the actual system transformation in order to simplify the changeover to the new system later on.

Process mining also supports you when it comes to master data. Smart algorithms and the use of artificial intelligence allow you to identify duplicates and find and clean up expired or incomplete master data. This allows you to tackle the issue of data governance directly.

System introduction: During the migration

An important aspect of a successful system implementation, especially for such complex projects as an ERP implementation, is a well-structured test phase. Extensive testing before the go-live allows existing problems to be identified and rectified in good time. In this way, you avoid the risk of disruptions in operation after the switch to the new system because certain functions do not work as intended or the customizing is not quite right. Process mining can be a great help for your test management, especially when it comes to functional tests.

The process analysis allows you to identify your most important business transactions. The test cases that are to be tested in the new system can be derived from this. By connecting the new system to the process mining environment, you can track the execution of the test cases and also monitor performance. Do the test cases run as planned or are there differences depending on the person executing them? If so, you may need to sharpen up the process training. Or if the required functions do not work, then it’s back to system customization. During the next test run, you can tackle the affected cases again and document the improvements made. This way, you always know where you stand, how much work still needs to be done and whether the overall project is still on schedule.

Adaptation phase: post-migration

However, the work is far from over after the migration and successful go-live. The majority of end users are only now starting to work with the new system. Here, it is important for the success of the transformation project that the new processes are also adapted and lived. Especially within the processes where a major change has taken place, it is important to ensure that the end users do not fall back into old patterns and work past the process.

Thanks to the data connection that was established during the migration phase, you have full transparency of your processes in the new system from day 1. You can quickly identify undesirable process variations or even be notified as soon as the system identifies a process variation that deviates from the target process. You can then use targeted training measures to accelerate the adaptation rate of the new processes and ensure the success of the project. In addition, monitoring process performance also allows you to evaluate which process improvements have resulted from the system transformation.

Living continuous improvement in the new system

The system transformation has been successful. The adaptation rate is high. The desired process improvements have been achieved. The successes were celebrated. The project team had a well-deserved breather.

Don’t rest on your laurels. Digital transformation is a continuous process. Today, process mining can do much more than just analyze processes. It is not without reason that providers now refer to their applications as an execution management system. In this sense, system transformation is the enabler that opens the way to other exciting projects. Here are a few examples:

1. Build smart alert systems that notify you of changes in your system that impact your operations.

2. Automate process execution, even across different systems.

3. Build dashboards tailored to specific business problems that provide you with all the information you need to make the best possible decisions. If desired, also with direct interaction in the system and supported by artificial intelligence.

You can also find more ideas in our consulting product and our publications on the subject of process mining.