What is process mining?
Process mining enables us to reconstruct, evaluate and visualise business processes across systems, using real-time data with the help of special process mining software. Modern mining algorithms create an interactive graph that allows deviations, loops or time delays in the process to be detected immediately and accurately – down to the smallest detail.
Event logs as data foundation
Process mining is possible thanks to the use of digital traces in IT systems. These tracks are typically captured by operating systems for every activity in a process and recorded in event logs. Classic logged data includes, for example, failed login attempts, system security warnings, or errors when configuring drivers.
Case study: a competence centre for process mining displays outstanding value
A 2022 study carried out by the Fraunhofer Institute in collaboration with software supplier Celonis demonstrated the advantages that professional process mining can offer companies. All the 214 customers surveyed stated that a "Centre of Excellence" (CoE) for process mining was a decisive factor for the rapid implementation of added value and efficiency gains. CoEs are competence centres responsible for the introduction of new technologies in companies, such as transmitting best practices, taking over technical implementation and offering support and training.
Process mining in purchasing: a successful step-by-step introduction
Process mining can also be of great help in procurement. Buyers need to keep an eye on their supply chain – the more transparent it is, the better. As the Deloitte CPO Survey 2023 shows, the chief product officers surveyed see their supply chains as being exposed to a wide variety of risks. In addition to high inflation, geopolitical uncertainties and the threat to continuity of supply are of particular concern. For almost two-thirds of the participants, an essential element of being able to minimize these risks is better transparency in the supply chain – including where upstream suppliers are concerned.
There is no universal recipe for success or firmly defined tools for the implementation of process mining. However, the following approach has proven to be effective:
Step 1: Determining the scope of the project
It all starts with planning. For this phase, it is appropriate to have workshops involving all participants. Here, the contents of the project and the process details are defined: which steps of the process should be analysed? What data is required for this? Should process mining be implemented internally with the help of tools, or externally, for example by a CoE, which can make practical recommendations for scaling process mining within the company?
Step 2: Data preparation/ process mining tools
In step 2, the technical requirements for process mining are created: how is the data extracted, transformed and transferred to the process mining software?
Two methods have proven effective for this: linking via a software connector and the use of an ETL (Extract, Transform-Load) tool. These typically pull information from data-based tables of transactional systems such as ERP or CRM, or from analytical data such as reports, log files or CSV files. Once this data has been successfully extracted, it is translated into so-called "cases": a sequence of distinct steps in process execution. All information about these cases is stored in the event logs accessed by the process mining software.
Step 3: Evaluating process data
Now the data can be analysed by process mining. Ideally, measurement should begin high up in the process flow, with the following phases of the process being analysed one at a time. To interpret individual data correctly, it may be necessary to interview those responsible for individual process steps.
Step 4: Measuring results
In the last step, possible improvements to the process are evaluated, tested and documented. The resulting changes are discussed by the team and implemented. Key performance indicators should also be consistently measured and monitored. It’s also advisable to extract new data after a while, in order to better identify what has changed and exactly which measures have resulted in greater efficiency.
Best practice: Grocery giant Ocado optimises processes and unlocks value with process mining
Ocado Group has a reputation for making the most of cutting-edge technology via its end-to-end commerce, fulfilment and logistics ‘smart platform’, which combines robotics, AI and machine learning. But with thousands of products and an average of 45 items in a typical grocery order, the company has numerous challenges to tackle. Streamlining processes for its growing number of international fulfilment centres is paramount. Partnering with software supplier Celonis offered the solution to Ocado’s challenge, given its ability to scale at the speed required.
Using process mining, the company has been able to troubleshoot many of its procurement issues, such as partners overriding AI-driven recommendations for purchase orders – which resulted in excess inventory being held, generating waste and increasing costs. In under a year, Ocado Group has realised value of at least 100 times the investment it made in Celonis’ process mining solution and intends to deploy it end-to-end across its own smart platform.
A 2020 research report by ABBYY indicated that 64% of UK businesses were already using process mining and that 91% believed it would be useful to their business. However, there’s still a lack of understanding in some quarters, coupled with a degree of scepticism about the technology.
And for the unconvinced, there is one compelling business case for process mining: its potential to protect against cyber threats. Process mining can proactively identify vulnerabilities in systems, through which attackers can penetrate, and it can detect potential loopholes more swiftly.
