Optimising business processes within your company is one of the most important factors for efficient operations. The process surveys were carried out using the traditional methodology of long interviews and documentation. This exercise was very time-consuming for both our consultants and the mapped company, and it was also problematic because the people involved often described reality according to their own experience, which led to contradictions and presented our experts with no small dilemmas in resolving them.
At KPMG we now have a much faster and more reliable approach. Process mining complements the traditional process analysis methodology, with the focus shifting to a data-driven approach. In a short interview, the company's business process experts describe which systems are involved in which process steps, and the log files of these solutions are the source of subsequent analysis. By analysing factual, time-stamped data from the associated databases, it is possible to map the interactions of the colleagues and departments responsible for the process, bottlenecks, redundant paths or time-consuming repetitions that can be eliminated.
To illustrate with a simple but very real example: in every company, generally accepted topos are developed. For example, "finance sits on payments for a long time". The stakeholders involved: the business area receiving supplier invoices, approvers and accountants all have their own opinions on this, but system data is perfectly objective. It may be that invoices are regularly submitted by colleagues with incorrect information or without a certificate of completion (invoices bounce back and forth, unnecessary repetition is introduced into the process), or that there is a necessary backlog of work in finance (permanent or intermittent bottlenecks), or that payments are only approved on certain days in the company because that is when the people responsible can dedicate time to reviewing invoices and certificates of completion.
The shorter the turnaround time and the more efficient the analysis, the less system data to work with and the more process steps to cover. It is important that each process or activity has an ID in each system, but it is not a problem if the same process runs under different IDs in each system, as long as this does not affect the existence of a link between process steps. It is also a condition that each system time stamps all actions. The process is system-independent in the sense that the analytical team can process data from any solution, be it an arbitrary ERP system (e.g. SAP), a CRM, or a custom-built target software, the point is that a link between them can be established through IDs, the rest is up to the analysts.
The shorter the turnaround time and the more efficient the analysis, the less system data to work with and the more process steps to cover. It is important that each process or activity has an ID in each system, but it is not a problem if the same process runs under different IDs in each system, as long as this does not affect the existence of a link between process steps. It is also a condition that each system time stamps all actions. The process is system-independent in the sense that the analytical team can process data from any solution, be it an arbitrary ERP system (e.g. SAP), a CRM, or a custom-built target software, the point is that a link between them can be established through IDs, the rest is up to the analysts.
In this way, we can get a quick and comprehensive picture of the business processes, or at least of the activities that are recorded in some system by the people involved. The use of process mining is the next step towards data-driven process improvement.
