I will always remember a particular phrase that my information systems science (ISS) professor once said: “If you think you can clearly demonstrate the value creation of any particular information system, you are probably wrong.”

This statement doesn’t mean that information systems don’t create value or that this value couldn’t be demonstrated. It means that measuring the value creation of implementing a system is extremely difficult.

However, process mining might just solve this issue.

How is value created within information systems?

To answer the question of how value is created within information systems, we have to consider what information systems are and what is their purpose. In general, Information systems are a combination of social and technical aspects. Social aspects refer to the people using the system and the way they use it, while technical aspects refer to the technology and processes used within the system.

In practice, we can imagine a silly scenario where I write a Post-It note, put it on the fridge and call it an information system. The people interacting with this system is me and whoever sees the Post-It, the structure refers to the way these people interact with the information of me as the sender and someone as the recipient, the technology used is a Post-It note, and a pen and the process is the transfer of data from me to the Post-It to anyone reading it.

To understand the value creation of this “Post-It information system” we have to consider the reason it was built. There can be an infinite amount of reasons why an information system is created and thus we have to ask the business owner (me) why I needed this elaborate system in place.

The business problem in our example was that someone was taking my lunch from the fridge and I needed to convey that this is in fact an error in the business process.

To resolve this issue, I placed a Post-It note to the fridge conveying that this problem should be addressed. Let’s say my lunch was disappearing 15% of the time before this system was in place and 5% of the time after the system was implemented. This process improvement is a measurable thing that proves that the system is working. Or is it?

The benefit of process mining over traditional process management

The Post-It system discussed earlier could be an example of a traditional approach to process management. The business owner in the case noticed a problem in a business process and implemented a solution aiming to resolve this issue. However, even after the system was implemented the root cause behind the process error was not solved. This is where process mining proves its superiority over traditional approaches.

Let’s imagine the business owner had collected data about the process before trying to improve it. In this scenario I take note of the circumstances every time the process occurs. By doing this I can notice patterns: the time my lunch disappears is always between 12.00 and 12.30, the day is Tuesday and so on. By analysing these “event logs” I can notice trends and do a root cause analysis.

It seems that the common factor between these data points is John from marketing. After discussing my suspicion that John is behind these incidents, the error rate drops to 0% and I can stop using my time on the Post-It system altogether.

Traditional process management can often be content with a seemingly successful process improvement without actually identifying the root cause behind the process failure. Modern process mining makes effective use of business data and creates business value through information systems more effectively than ever before.

To conclude, using modern process mining techniques can solve the age-old problem of creating tangible business value through IT investments. The future of process improvement is solving the root cause of the problem and not just muddling towards some arbitrary KPI level.

Read more here

 

About the author

Valtteri Nättiaho is part of Sofigate’s Data Leadership team, where he helps customers tackle the challenges of reporting and data analytics.