Process mining and ERP: two sides of the same coin

Process Mining ERP - GUS-OS Suite - GUS ERP

There are now more than 35 operating tool providers for process mining. And according to Fortune Business Insights, the global market is expected to grow from USD 1.66 billion in 2023 to USD 27.72 billion in 2030. This corresponds to an annual growth rate of 49.4 percent during this period. Process mining is becoming a standard discipline. And more and more companies using ERP solutions are realizing that their ERP system can serve as a central data source for process mining.

By 2009 at the latest, process mining had become an independent discipline in the world. At that time, the global professional association IEEE (Institute of Electrical and Electronics Engineers) set up a task force on process mining. Since then, the committee has been working on norms and standards in connection with process mining, for example in the current "Process Mining Manifesto" from 2011.

Process mining is a data mining technique, but goes beyond traditional data mining by enabling the systematic, data-supported evaluation of entire business processes. The core of the technology consists of analyzing the digital traces of the processes - as provided by the respective IT systems in the form of log and event data, for example. Business processes can be reconstructed, monitored, examined and optimized with the help of process mining. Process mining is able to combine and visualize individual activities into an overall process using such process data. The great strength of this approach is that it is based on real data from live operations. This creates an objective and up-to-date picture of the business processes under investigation. Differences between the actual existing processes and the target status on paper become visible.

Recognize, check, improve

Since the publication of the IEEE manifesto at the latest, a fundamental distinction has been made between three different types of process mining: "discovery", "conformance" and "enhancement": 

  • The first type is discovery. Here, log data from events is used as input to generate a process model without the need for further information. Additional data such as organizational, meta or other context data can help to further refine the derived process models. In principle, however, the usual event logs are sufficient.  
  • The second form of process mining is conformance checking. This procedure examines and documents the delta that exists between the defined process standards and the actual log data. In this respect, this check shows the extent to which the reality documented in the log data actually matches the existing models - or not. For example, process models, organizational charts, declarative models, business rules, guidelines or similar can serve as reference models.  
  • The third type of process mining is enhancement. The aim here is not only to identify and analyze deviations between actual and target processes, but also to expand and improve the existing process model using the information from the event data. If, for example, the time stamps reveal bottlenecks or excessively long throughput times, those responsible can change the process in order to eliminate the localized "bottlenecks".

Lower costs, higher productivity

Companies primarily use these three methods to analyze and optimize their business processes. However, this is also only one means of achieving further-reaching business benefits. First and foremost is the reduction of operating costs. This can be achieved by identifying inefficient processes, automating tasks and reducing waste. Process mining also helps companies to increase productivity by providing insight into the actual work processes taking place, identifying and eliminating bottlenecks and enabling employees to work more efficiently.

Process mining can also help ensure that processes comply with legal and regulatory requirements. This is particularly important in highly regulated sectors such as healthcare or the chemical and food industries. Other benefits include improved risk management thanks to early detection of problems with the help of process analysis, greater data transparency, particularly for data-based business models, and optimized use of resources, for example to meet sustainability targets. Last but not least, process mining can increase customer orientation, as it enables companies to view their processes from the customer's perspective.  

What characterizes good process mining tools

The market for process mining tools is one of the most dynamic software markets of all. Since the first commercial process mining provider ("Futura Pi") was founded in 2007, dozens of other companies have attempted to position themselves as providers. Today, there are more than 35 tool providers on the market worldwide. What features characterize the best tools?

Leading process mining tools offer data import from various sources and formats, including log files, databases and other enterprise systems. They are also able to process large volumes of data and facilitate the integration of data sources. At the same time, good process mining tools enable data cleansing and pre-processing to remove "noisy" or incomplete data and prepare it for analysis.

User-friendly visualization is crucial for presenting complex process flows in an understandable way. The best tools offer a variety of diagrams, dashboards and heat maps that make it easy to visualize the process flow, bottlenecks and deviations. The core task of the tools also includes process analysis functions to uncover bottlenecks, throughput times and efficiency problems, including conformity checks: does the observed process comply with the specified process models or rules?

Advanced tools often offer predictive analytics capabilities to forecast future process trajectories and issues. It should also be possible to export results to other analysis tools or business applications so that the results can be translated into action. In addition, features such as scalability are required in order to be able to process large volumes of data efficiently and carry out sophisticated analyses. Customization options, in turn, make it possible to take company-specific requirements and business processes into account. Last but not least, process mining tools should offer robust security and data protection functions, as they often deal with sensitive company data. And good customer support and extensive training options are important to ensure efficient use of the software.

Current market trends

In addition, the various solution providers differ in terms of their ability to keep pace with current developments in the process mining tools market. All leading process mining tools are now increasingly relying on advanced data analysis techniques and artificial intelligence (AI). These include, for example, predictive analytics, machine learning and automated recommendations for process optimization.

Cloud-based solutions are almost standard, as they improve scalability, flexibility and collaboration in distributed teams. Cloud-based process mining tools also enable companies to access resources quickly without having to maintain their own extensive IT infrastructure.

Solution providers have recently made great progress with process discovery functions that automatically identify unknown or poorly documented processes. The linking of process mining and RPA is also becoming increasingly important. Companies are using process mining tools to identify potential automation opportunities and then use RPA bots to automate manual tasks.

While traditional process mining tools analyze historical data, real-time process mining solutions have also recently emerged. With their help, companies can analyze process data in real time to gain immediate insights into current processes. The user group has expanded in recent years thanks to low-code or no-code approaches. This means that even business users without in-depth technical knowledge are now able to carry out process mining analyses and optimize processes.

As in the ERP world, the motto "one size fits all" is becoming less and less valid for process mining. Many providers have recently been particularly successful by developing and launching industry-specific process mining solutions. As a rule, they are better able to address the specific requirements and challenges of individual industries.

ERP data and process center

Most (manufacturing) companies have a major advantage when it comes to process mining, as they do not have to start from scratch thanks to their already established ERP system. When introducing and maintaining an ERP system, those responsible necessarily think about their processes. This includes, for example, process visualizations or making the different (subjective) opinions of the stakeholders involved in the company transparent. It is true that these subjective views often have little to do with what actually happens in the company. But at least there is already an analysis of the business processes.

This analysis includes simple but essential questions: What business processes and objects, such as production orders or purchase orders, are there? What stages do these processes go through? And where in the existing systems does the ERP already collect data on these business objects and activities? This data input from the ERP on manual or automated processes is a valuable basis on which processes can be identified and further developed with the help of process mining - and all this from within ongoing operations.

Without this constantly updated ERP data, process mining runs the risk of lagging behind reality. Once the process snapshot of the processes and their visualization has been taken and those responsible for business process modeling in the specialist departments are on board, there is a risk that the process has already evolved in reality and the picture drawn is already outdated again. This is where an ERP can provide objective, up-to-date and reliable data.

ERP solutions also provide valuable information for process mining when checking the conformity of processes. In the ERP solution GUS-OS Suite, for example, it is possible to define business objects along such workflows. These instantiated workflows can be used to check which processes really exist and which only exist on paper or in the minds of employees. Ultimately, the ERP performs a reality check: Are processes being lived as they should be, or are certain user interactions regularly skipped or bypassed? In short: Has the exception become the rule?

Real process data instead of subjective assessments

When it comes to process improvement, ERP systems can provide important key figures. For example, for key questions such as: At which points do certain tasks always remain unfinished for a particularly long time and for how long exactly? How long is the lead time from the placement of an order by the customer to the delivery of the order or to the point at which the customer pays their invoice? This is where an ERP system provides reliable real data instead of subjective assessments by individual stakeholders - be it the head of department or the managing director.

But not all ERP systems are the same. With regard to process mining in particular, there is a big difference between workflow and process-oriented solutions and those that do not offer this structure. A second important distinguishing feature is the ability to present central key figures in the form of a process and data cockpit, preferably in real time. The ERP is then already able to show whether and by how much, for example, the order throughput has become faster or slower in the previous month. Process mining can then be based on such grouped and weighted data.

The same applies to data that has not yet been collected - for example, supplier orders that are still approved manually and therefore do not leave a digital data trail at this point. Such steps, which are still "paper-based", are predestined to be left lying around for longer or to contain careless errors. An ERP system can already identify such gaps in the workflow. This enables those responsible to automate such "manual islands" and then measure the extent to which the order throughput has accelerated. This makes it possible to identify precisely which digitization steps save the most resources and bring the greatest added value. Process mining can then start at those points that optimize the processes in a targeted manner.

Appropriate reporting, which summarizes, adds up and then displays relevant data and even carries out certain evaluations, is therefore also part of many ERP systems - such as the GUS Report component within the GUS-OS Suite. The relevant data can be further "refined" here using filters or restrictions. The data linked to the processes should then also be visualized in the ERP - usually in the form of dashboards, which should be as freely definable as possible.

Process industries at an advantage  

Companies in the process industry generally have a certain initial advantage when it comes to process mining, as they already operate in a highly regulated environment. This primarily includes the pharmaceutical, medical technology, chemical, food and cosmetics industries. Due to the numerous regulations, laws and other industry requirements, their processes must be transparent and regularly validated. In these industries, process mining is primarily used to further increase the level of process maturity in order to gain a competitive advantage. In these companies in particular, the ERP system must also be process-oriented so that it can map the respective company processes without detours and provide reasonable support for process mining.

Fact-based decisions

ERP and process mining solutions therefore complement each other perfectly for various reasons. On the one hand, process mining has the functionalities that can draw conclusions about business processes from system data and subsequently improve them. On the other hand, an ERP system such as the GUS-OS Suite always provides up-to-date data that originates from the real existing workflows in the company. Process mining is therefore able to draw qualified conclusions about the maturity level of the processes and identify inefficiencies and optimization potential. The result is cross-company process controlling based on real-time data. Those responsible no longer have to rely on processes that in some cases only exist on the drawing board. Instead, they are able to make fact-based and reality-based decisions based on their actual and current business processes.

FAQs: Frequently asked questions about process mining

What is process mining?

Process mining is an approach that is positioned between computer-aided intelligence and data mining on the one hand and process modeling and analysis on the other. The idea and aim of process mining is to discover, monitor and improve actual (and not just theoretically defined) processes. This is done primarily by extracting knowledge from event logs that are already available in today's (information) systems. Process mining includes the (automatic) discovery of actual processes based on event logs and checking whether model and actual processes conform or differ. Derived from this, optimized process models are automatically created, expanded, maintained and process predictions are made on the basis of historical system data.  

What are the advantages of process mining?

Process mining is able to extract knowledge from event logs of today's information systems. This enables the discovery, monitoring and improvement of processes in a variety of application areas. This in turn improves, controls, accelerates and streamlines business processes especially, but not only, in highly competitive and rapidly changing environments/markets.  

What data does process mining require?

The data required varies depending on the objectives and scope of the process mining analysis, but in general the following types of data are required:  

  • Event logs containing records of all activities and events that occur in a business process. Each event in the log should contain at least the following information: Timestamp, case or process instance ID, activity name or ID, resource or personnel ID and other relevant attributes such as product/customer number, location, etc.
  • Organizational data with information about the organizational structure, resources, roles and responsibilities within a company.
  • Reference data that makes the process context understandable and thus allows the real process to be compared with the desired processes. Reference data can be historical data, benchmarks or other relevant information that helps to identify deviations or optimization opportunities in the process.
  • Additional context data that may be required depending on the specific requirements of the process mining analysis - for example, customer, product and supplier information or other relevant data needed to analyze the process.
  • Metadata with information about the structure of the event logs, such as field names or data types. They are important in order to interpret and process the data correctly.

In which industries can process mining be used?

In general, process mining can be used in any industry in which business processes exist. It enables companies to bring transparency to their processes, identify efficiency potential, ensure compliance and improve the quality of services or products.

How does process mining differ from data mining?  

Both data mining and process mining fall under the heading of business intelligence. Both use algorithms to understand big data and can also use machine learning. Both can help companies to improve their performance. However, process mining is more about how information is generated and how it fits into entire business processes, whereas data mining relies solely on the data available. Data mining is therefore more concerned with the "what", i.e. the patterns themselves, while process mining looks for the "why". In this respect, process mining goes beyond data mining, as it also uses data to uncover and eliminate differences between desired and actual business processes.  

Can I integrate process mining into my existing IT systems?  

Yes, but the successful integration of process mining requires comprehensive collaboration and planning between process management, data analysis and the IT department. The following integration steps must be carried out: Identifying the business processes to be analyzed; preparing the required event logs/data sources; checking the IT infrastructure for integration options for process mining tools; selecting and implementing a suitable process mining tool; training the responsible teams; importing, cleansing and transforming the data into and for the tool; analyzing and improving the business processes based on this and continuously monitoring them.

Can process mining be optimized by using an ERP system?

Definitely. An ERP system provides valuable data input on manual or automated processes in a company. This data can be used to discover, analyze and further develop business processes through process mining. In process-oriented ERP solutions, it is possible to define business objects along workflows. Here, an ERP system offers reliable real data instead of subjective assessments by individual stakeholders.

How secure is my ERP data when using process mining (software)?

When it comes to the security/backup of ERP data in process mining, the factors that also apply to communication between other software systems essentially apply. These primarily include the security level of the software itself, its implementation, access control, data encryption, the type of installation (local or in the cloud), the possibility of data recovery through backups or adherence to data protection regulations (compliance).  

How can the GUS-OS Suite as an ERP solution support process mining?    

Above all, the GUS-OS Suite provides process mining with data that is always up-to-date and based on process reality and prepares it in the form of reports and dashboards. This means that initial analyses can already be carried out within the ERP with a view to process improvements. Further functions using the ERP data are then offered by process mining itself.

For small and medium-sized companies, an ERP system such as the GUS-OS Suite is often the central data hub in the company and is therefore also the most important data source for process mining. Larger companies, on the other hand, usually have several IT systems in use, such as for customer relationship management (CRM). In this case, it makes sense for the respective systems to either be able to access each other's data. Or a selected system acts as an overarching instance in which all essential data flows together. This presupposes that - as is possible with the GUS-OS Suite - the other systems involved can also be integrated with each other. This is usually done via so-called REST interfaces.

Can process mining software analyze real-time data from my ERP system?

An ERP system generates data on business objects such as a purchase order or a sales order as well as on transaction data such as stock movements and value flows at the moment they are entered and processed. If this is accompanied by the generation of a time stamp, process mining software is able to analyze this data and incorporate it into overall considerations. If the ERP system then provides additional runtime information on these processes on the basis of workflows, the process mining tool is also able to react to short-term deviations in the process chain and make them visible to the user. By comparing the current situation and the shorter and longer past data, targeted conclusions can then be drawn for process optimization.

Get into conversation

Play video

Newsletter - GUS Insider

Don't miss any more news and sign up for the GUS Insider. Here you will be regularly informed about all news of the GUS Group and about our software.