Process Mining – Business Process Management’s Digital Future?

For many companies their own processes are an impenetrable black box. That is why business process management has for decades sought to shed light on the subject. Answering the question who does what, where, when, how and, above all, to what purpose within the company is usually complicated, protracted and subjective. Real truths can now be established digitally on the basis of digital facts by using process mining technology. But does this innovation really have the potential to turn classical business process management on its head? We discussed with Florian Umbreit and Tim Bodenstab, consultants at mgm consulting partners, their personal opinions of process mining.

mgm live: Florian, Tim, you deal with business process management issues on a day-to-day working basis. Why should companies deal with the subject at all?

Tim Bodenstab: Most companies have long realized that business process management is necessary. Many firms repeatedly learn the hard way that their processes are not running smoothly. That is only natural. Especially at large companies many different people, departments or even country companies are often involved in the different processes. That is why the painful experience that there are problems with their processes is one of the most frequent reasons why companies want to look into process management. They would like to put their processes to the test and ideally to optimize them.

Another motive can be that organizations want or need to document their processes in general. Against the background of aging workforces and a shortage of skilled workers there is a serious risk in many companies that the undocumented knowledge employees have in their heads will be lost when they retire. This development is a serious risk for the business continuity. Furthermore, pressure on margins or other external factors that pose a risk for companies can prompt them to look into their business processes and mitigate this pressure by means of greater efficiency.

mgm live: So are companies open to the subject?

Florian Umbreit: Our day-to-day working experience is that business process management has a very hard time of it in companies despite its enormous potential. There are various reasons why that is the case. High-quality, sustainable process management in the classical sense of the term is, for example, complex and labor-intensive. You often meet with different contacts from departments for months to discuss and eventually find out how a process actually works. This frequently leads to an overview of a handful of process variants that at times greatly depends on the views and perceptions of participants. So you have no guarantee that the variants identified reflect the objective truth.

The problem is compounded by the fact that business process management is one of the disciplines of which the success is very hard to measure. Even if optimization impulses result from process management they do not by themselves change anything. In order to generate a perceptible value they must lead to specific implementation projects. If these are a success but process management is not sold to the Management as the originator and driver of these improvements or the success cannot be described in detail because it is difficult to measure, this management method usually appears to be less useful than it is and gains a bad reputation.

mgm live: Process mining has for some time been establishing itself as a new business process management technology. How does it work and what makes it different from classical methods?

Tim Bodenstab: Process mining makes it possible to analyze business processes digitally. The software makes use of the fact that processes today for the most part run on IT systems. The individual process steps leave digital traces in these systems in the form of a process ID, activities and a time stamp. The process mining software finds these traces, evaluates them and thereby generates an exact map of the process cycle that includes all variants. The company then has a dynamic real-time documentation of the processes that provides information about both how much time elapses between individual process steps and the different ways in which the process is handled in reality.

mgm live: Is acquiring the necessary data not a problem? In large enterprises it is often stored in different systems.

Tim Bodenstab: Extracting data is indeed a challenge—and a reason why the implementation of process mining software can take time, depending on the individualization of IT systems. You need to understand the process relevance of individual databases and to be able to gain an idea of where the traces of individual process steps are stored. This phase requires a great deal of brainpower. In our experience, at least in collaboration with our software partner Celonis, it is much faster with standardized ERP systems such as SAP, ServiceNow, Salesforce, etc. because they often rely on standards in process execution. But in principle any ERP system or process-relevant database can be connected to process mining software. Linking IT systems was further simplified by using the Celonis Intelligent Business Cloud.

mgm live: Where do you see specific advantages of process mining over classical business process management?

The crucial advantage of process mining is that the documentation of processes is based on facts. In this way decision makers receive a reliable picture.

Florian Umbreit: The crucial advantage of process mining is that the documentation of processes is based on facts and does not reflect the assumptions and opinions of individuals. In this way decision makers receive a reliable picture on the basis of which they can decide which processes are to be looked at in greater detail and may be operatively optimized. In all honesty the corporate reality is often that the project or change budget goes to the person who shouts loudest. With process mining you can buck that trend and argue on the basis of facts. “You may have shouted loudest but what you want to change is negligible in relation to the totality of the process because it only accounts for two percent of cases or the volume of business.”

A further advantage is the bandwidth of process variants that can be analyzed by means of process mining. Rather than the four or five variants that come to the fore in classical business process management discussions you get the full truth about the individual process paths. Which of them was I expecting, which of them are unusual? Which are harmful or non-compliant? At this point in process mining you can filter precisely and limit peculiarities further and further. Instead of discussing for weeks or months you can use digital traces of the actual work that exist in any case. All these advantages address at the same time the disadvantages of traditional business process management so that process mining is a very helpful addition to the classical methods.

mgm live: Are there limits to process mining?

Tim Bodenstab: In the nature of things situations in which an employee discusses a process step with a colleague over a cup of coffee leave no digital traces in IT systems. So the software cannot register them. Nevertheless, even corridor chats of this kind can leave indirect traces. Process steps that require regular personal consultations of this kind may take an unusually long time. If analysis of the findings reveals long process times it might make sense to ask the process participants why that is the case. Process mining after all only indicates points that might benefit from further investigation into why certain peculiarities occur. The causes are in part only apparent from the context.

It is at this point that a further limit to process mining—at least in Germany—comes to light. If you want to use the technology to check process paths by employee you will soon come up against data protection and corporate co-determination. In Germany, fortunately from the employees’ viewpoint, individual employees’ performance can only be compared by means of process mining with the consent and involvement of the Works Council. In other countries, such as the United States, that is not the case.

Florian Umbreit: We must also always bear in mind that linking the software is not enough. Buying new software doesn’t mean anything is going to change in the company. There must be a readiness to deal intensively and in the long term with one’s own processes and mistakes and to implement the findings in earnest. If I just run the software on the side and would otherwise prefer to keep to the tried and tested paths the investment will not be worthwhile.

mgm live: Process mining is still a very young technology. Where do we go from here, would you say?

The technology certainly still has some way to go developmentwise. My personal forecast is that the software will become more and more independent in the years ahead.

Tim Bodenstab: The technology certainly still has some way to go developmentwise. My personal forecast is that the software will become more and more independent in the years ahead. I can well imagine an increasing incorporation of artificial intelligence and machine learning leading to the software recognizing patterns itself and telling the user “there’s a discrepancy here you should take a closer look at”. The development of Celonis’s Intelligent Business Cloud is heading in that direction and providing good approaches to automated process analysis.

mgm live: Does that mean human process managers and consultants will at some time or other cease to exist?

Florian Umbreit: No-one can say for sure where development will end. For the foreseeable future humans will definitely continue to play a central role in business process management, however. The software, as noted earlier, merely suggests points at which more detailed investigation might be advisable. For substantial analysis of peculiarities and their evaluation, however, personal contact with the process participants will continue to be necessary. In some work situations humans still intuitively opts for the better approach because they are aware of the wider consequences of what they do. Sometimes, indeed, it makes sense to “bend” a process deliberately in order to make an irate large customer happy again.

That too is why I am not pessimistic about the future for us as management consultants. For one, we can of course contribute our methodical know-how to the launch and operation of project software. For another, process optimization only gets under way after the analysis undertaken by the software. When you consider, for example, the change management component of transformation projects of which the optimization forms a part, I feel sure that business process management will continue not to be able to get by without people.