“Sometimes process owners are surprised by their own data.”

As a specialist in process mining, Henny is responsible for the technical implementation of Signavio Process Intelligence, and helping customers with process analysis. Henny holds a Masters in Data Science, and is experienced in the implementation of IT transformation projects. In the following interview, she’ll tell us what Process Mining initiatives are all about.

Process Mining expert - Henny Selig

Henny, as a consultant you support our customers with extensive process mining projects, as well as with implementing Signavio technology. How would you explain the concept of process mining to novices?

Process mining is a combination of data analysis and business process management. Many processes happen automatically or are supported by IT systems, meaning they leave behind digital traces. We extrapolate all the data from the supporting IT systems connected to a particular process. This data is then visualized and evaluated with the help of data science technology.

Why is process mining so important for process owners?

Process analysis which is supported by data enables a fact-based discussion. Therefore process mining builds a bridge between employees, process experts and management. This helps avoid siloed thinking, as well as enabling transparent design of handovers and process steps which cross departmental boundaries within an organization.

Which use cases are particularly suited to data supported process analysis?

Every organization is different and brings with it a variety of process-related questions. Yet some patterns are usually repeated. For example, we have customers who introduce data supported process analysis as part of a business transformation. The challenge with such a project lies in harmonizing processes from fragmented sectors and regional locations. Here it helps enormously to base actions on data and statistics from the respective processes, instead of relying on the instincts and estimations of individuals.

Sometimes customers want to optimize their processes in line with a particular goal, to improve customer experience or cash flow for example. Other organizations need a complete overview of the actual workings of processes as part of company strategy. In this case organizations are normally interested in continuous improvement. Often, even processes that are initially seen to be efficient can be optimized further.

In such projects, do the findings mostly correspond with expectations?

It can happen that process owners are surprised by their own data, for example, if they overestimate their process performance beforehand. In this case it’s helpful to examine the reasons for the discrepancies in process behavior together with the potential for optimization. The solutions are usually easier than you’d think. In one scenario for example, simply changing the priority of certain activities already made quite a difference!

How can you test whether or not a predefined process is actually running as planned?

With process mining technology, I can take data from the process in its actual state and directly compare it with the modeled process, which is known as carrying out a target-state/current-state comparison. However, there is more to process conformity than carrying out process steps properly. It is also important to ask:

  • Were the limits for approval or discount taken into consideration?
  • Were all required stakeholder approvals collected from various departments?
  • Is the timeframe realistic or are efforts distributed other than planned?

I can check for all these things with the data in Signavio Process Intelligence analysis module by having a look at all process variations. These variations happen for different reasons. For example, sometimes the target state model is defined in overly simple terms and therefore isn’t realistic. In other cases, there is a lack of process knowledge and IT support.

What should process owners be aware of when assessing these variations?

Not every variation that is out of line with the target model is necessarily negative. Very few processes, apart from those that run entirely automatically, actually conform 100% to the intended process model—even when the environment is ideal. For this reason there will always be exceptions which require a different approach. This is the challenge in projects: Finding out with all stakeholders which variations are desirable and where necessary exceptions must be made.

So would you say that data-based process analysis is a team effort?

Absolutely! In every phase of a process mining project a variety of project members are included. IT makes the data available and helps with the interpretation of the data. Analysts then carry out the analysis and discuss the anomalies they find with IT, the process owners, and experts from the respective departments. Sometimes there are good reasons which explain the way a process is behaving, which require explanation by an expert.

In this discussion it is incredibly helpful to document the thought process of the team with technical means, such as Signavio Process Intelligence. In this way it is possible to break down the analysis into individual processes and to bring the right person into the discussion at the right point without losing the thread of the discussion. Then, the next colleague who picks up the topic can then see the thread of the analysis and properly classify the results.

To find out more about how process mining can help you understand and optimize your business processes, visit the Process Intelligence product page. Or, if you’d like to get a group effort started in your organization right away, sign up now for a free 30-day trial with Signavio. Finally, stay tuned for a new Signavio white paper coming soon, which will expand on Henny’s insights, and offer even more more ways to help you make better decisions, faster.

Published on: March 5th 2018 - Last modified: April 27th, 2018