In a corporate context, many enterprise processes are partially or even fully supported by IT systems: the digitalization of processes represents more and more activities, supported by a rising number of systems that generate ever more data.

That being said, it is legitimate to ask whether or not traditional ways of learning processes are still ample:

Is documenting a vision of the goal process sufficient for the process to be applied in follow?

When a deviation from a model is perceived, is it optimal to seek consensus in a gaggle from subjective factors of view?

Is it possible to measure the precise execution speed of the process from start to finish?

Process Mining provides a new approach to take these parts into account.

A primary definition

Process Mining is an analytical approach that goals to build an exhaustive and goal vision of processes based on factual data.

Thus, Process Mining is a high worth-added approach when it comes to building a viewpoint on the precise implementation of a process and identifying deviations from the perfect process, bottlenecks and potential process optimizations.

How does it work?

Regardless of the nature of the process , as quickly as it is supported by digital tools, information is created and stored by the corresponding IT systems (ERP, enterprise applications, etc.), in particular through application logs. This stored information often has comparableities and makes it possible to hint the path of an «object» by way of different levels at completely different times in time.

Process Mining is predicated on instruments that use these digital footprints to reconstruct, visualize and analyze processes, thus providing transparency and objectivity towards the real process.

Required data

With the intention to be usable, these digital footprints must no less than embrace:

Object: an instance that will be followed throughout the process, with a novel identifier. The choice of this object influences the scope of the studied process

Activity: a step within the studied process. The selection of activities influences the granularity of the process

Date: determines the order of activities and timing

In addition, it may be fascinating to collect additional data relying on the process, for instance: supplier, type of product, location, person/administration, channel, worth…. These will enable further investigation.

Process visualization and analysis

From these data, it is possible to visualize a illustration of the ideal process and all deviations from it. This allows for early detection of potential inefficiencies within the process.

Past the illustration of the process, one can even look on the execution instances of each step, or look at a more limited scope to be able to establish where, when and why the process deviates from its ultimate version.

Instance with a purchasing process

For a simplified purchasing process ideally composed of 4 steps («Record the order», «Obtain the products», «Record the bill» and «Pay the bill»), the process followed by orders is traced from the digital footprints left in an ERP.

Use cases and benefits

There are three major use cases of Process Mining:

Discovery: building a vision of an existing process when no model exists a priori

Verification of the correct implementation and analysis of deviations from a previous model

Process improvement

In all three cases, it is the understanding of the actual implementation of processes, primarily based on goal and exhaustive data, that makes the added value of the Process Mining approach.

In addition, this approach represents an improvement in the subject of process management:

Acceleration of studies (limitation of time spent and number of interviews) to build a illustration of current processes

Taking into account more data, or even the exhaustiveness of data, in the measurements

Opportunity, as soon as a new process is designed, to ensure effective management of its use and to see improvements

Process Mining is not dedicated to a particular sector of activity: the approach will be able to carry value wherever processes are applied and studied. Within an organization, a number of functions could also be interested within the approach:

Operational excellence teams: complementing the methods already used (Lean, Six Sigma, etc.)

Data Scientists: visual representations of data to generate new insights

Process managers: factual analyses to enhance their knowledgeable vision

CIO: vision of the usage of the systems and the corresponding person paths

Audit or internal management: faster evaluation and the possibility of relying on the exhaustiveness of cases moderately than on a pattern

Key success factors

With a view to obtain good outcomes, the launch of a Process Mining initiative requires some precautions. It may be noted that it is necessary:

To determine from the outset the added worth goal: cost reduction, improvement of the consumer/customer expertise….

To define a well-defined research scope when it comes to process

To operate iteratively with brief cycle analyses, within a fixed total time limit

To ensure the quality of the data on which the examine is based. To do this, it is essential to collaborate with the IT experts of the systems used as well as the enterprise experts of the processes studied

To accompany the change in case of redefinition of a goal process

Moreover, the analyses carried out by Process Mining should not be an end in itself but should function a factual starting level for additional process studies. Reintroducing a human facet, for instance by using a Design Thinking approach, makes it attainable to deepen the outcomes obtained thanks to Process Mining by taking the end users into account.