Organizations are drowning in data. The challenge is knowing how to exploit these data treasures by extracting valuable information from them. Process mining refers to a family of analytical techniques that can be used to discover, monitor and improve processes, based on the data extracted from systems such as ERP. One of the first questions we ask when meeting our clients is “What is the key process that people are really interested in and follow? What do they do in practice?” After all, what they tell they do might be different from what they actually do. That’s exactly the question we try to find answers to – what do they really do? What are the bottlenecks? Where are they? What causes them?
Once we know the process of interest, we initiate the data extraction and data engineering tasks. The required data needs to be extracted from the relevant systems, with the help of the client. The extraction is followed by a great deal of data engineering on our side. Typically, we deal with massive amounts of data and construct event logs detailing the course of events in specified processes. For this, we need to have precise knowledge of the client’s systems and processes, as well as skills in big data analytics. We are currently building an automated environment for process mining at KPMG Lighthouse, collecting the best tools and practices together. After the data engineering part, the clients’ data is ready to be loaded and explored for process mining.
Two examples of processes we have recently mined are the Purchase to Pay (P2P) and Order To Cash (O2C) processes for industrial clients. In general, our clients’ data sources support process mining. Yet, we usually find areas for improvement.
In these assignments we identify the “algorithmic happy path”, which is the most efficient scenario for the most frequent activities and process connections. After that, we explore the deviations from the happy path, and the root-causes of these deviations. For instance, we may find that, for specific vendors, there are high throughput times that only occur with the same purchase item groups. Currently, one of our clients has planned actions to mitigate such processes that are either slowing down the overall processes, or are causing repetitive manual tasks. Further concrete action points and the extension of process mining to cover other processes are currently ongoing.
One of the main advantages of process mining is the transparency it provides regarding end-to-end processes – a transparency that may not even be possible through traditional means. Understanding the entire process is a pre-requisite for transforming it.
Furthermore, we are also able to carry out continuous monitoring of a process and trigger automated actions based on the data. This enables firms to react immediately to issues in their business environments, which may even contribute to their competitive edge.
With process mining we help our clients to discover what they actually do, instead of what they think they do. As a result, we can identify those process areas that carry risks, and recommend improvements. Moreover, process mining is typically an iterative process that opens up opportunities for new analyses. For example, the identification of repetitive manual tasks and their influence on the overall process may elicit discussions on “how we can automate” such tasks. Ultimately, process mining is not just a means of analyzing historical event logs, but is also a way of developing the capability to predict where process deviations might occur. Such operational support can be implemented with machine learning expertise and integrated into the client’s environment.