Tricky point with Process Miningacadariu2019-02-11T14:16:59+00:00
The tricky point with Process Mining lies in the transformation of the data.
Process Mining can only be used successfully if the seamless connection of the data sources and the continuous transformation of the raw data is successful. This is exactly where the wheat separates from the chaff in the Process Mining tools.
“That’s not possible” was the typical reaction when I presented Process Mining to the first interested parties in 2009. Hardly anyone could imagine at that time that it could be possible to reconstruct processes from digital traces of IT systems alone.
ING-DiBa alone and its then Head of Internal Audit, Volker Panreck, were innovative enough to dare a first pilot in the field of construction financing processes. The deployment was extremely successful, Process Mining was introduced and later the bank was awarded the “Audit Innovation Award” at the Frankfurt School of Finance and Management, which also gave the starting signal for a broader use of Process Mining in business.
The rest is history and a lot has happened since then. However, for many managers, Process Mining is still an almost unknown topic. But numerous successful implementations prove that it works and provides benefits. Nevertheless, there is always a certain scepticism during presentations, which does not question the technical function of Process Mining, but the effort of data preparation and integration. And asking that question is more than correct. Only if the seamless connection of data sources and the continuous transformation of raw data is successful can process mining be used successfully in the long term. And this is exactly where the wheat separates from the chaff in Process Mining tools.
Which data is needed for Process Mining at all?
To be able to reconstruct a process with Process Mining, so-called event logs with at least three attributes are required:
a unique transaction number (for example, invoice number),
a time stamp (date and time) and
The activity executed (for example, release).
From digital footprints to process
In addition, other attributes such as supplier names, invoice amounts or persons involved in the process help with the further analysis of the processes. Systems such as SAP generate all the required data, but do not store it in a directly usable event log. Instead, they distribute it in numerous database tables.
How does the data access and transformation take place?
Almost all Process Mining tools on the market are limited to the visualization and analysis of processes. The data required for this is extracted from the databases of SAP or other ERP systems, for example, using third-party tools, then imported manually or occasionally semi-automatically into relational databases and then converted into the event logs described above using individually programmed SQL procedures. Occasionally these collections of SQL procedures are also called connectors, which is a pretty marketing term, but insufficiently masks the high effort, the possible sources of error at the interfaces and the unnecessary duplication of data.
However, Process Mining always unfolds its greatest benefit where it is not only used for one-off snapshots of processes, but also for their continuous analysis and optimization. The prerequisite for this is a data transformation that is as integrated and high-performance as possible, as it is contained in ProcessGold’s Process Mining software.
As usual with professional Business Intelligence platforms, ProcessGold has real connectors (interfaces) that allow direct connection to any relational database (HANA, MS SQL, Oracle, …) and file format (CSV, XML, XLS, …). The subsequent transformation, i.e. the conversion of data from different sources or tables, also takes place within the ProcessGold platform, using a proprietary in-memory database.
From data to insight
This way of data connection and data transformation is not only particularly high-performance, but also enables the use of Process Mining for topics such as Process Monitoring and Continuous Auditing.
How does the visualization and analysis of the processes take place?
Since ProcessGold was founded more than nine years ago, several hundred projects have been carried out worldwide and it was not until the end of 2016 that the company’s own Process Mining platform was launched on the market. All our experience flowed into the development of the “ApplicationOne”. This pre-configured solution enables a quick entry into Process Mining and allows the immediate analysis of all business processes in the company. Users gain valuable insights into how their processes actually work and where they can be improved.
At the same time, the ApplicationOne also represents a template that can be extended and adapted to individual wishes and needs as the project progresses.
Unlimited configuration options
Rudolf Kuhn has been working as an IT and process consultant and entrepreneur for over 20 years. In 2010, he founded ProcessGold AG in Frankfurt am Main, which was the first company to deal with the commercial use and consulting of Process Mining. After a merger with the Dutch software provider MagnaView, ProcessGold released its own Process Mining solution at the end of 2016, which, for example, is used by Ernst & Young as the standard in auditing by hundreds of clients worldwide.