Process Mining Basics

Process Mining? Data Mining? Business Intelligence? There are many analytical terms and it can be challenging to understand them all. This course will help you understand the terms associated with Process Mining and how it can be used to improve business processes. 

Before we dive into the world of Process Analytics, let's first understand: 

• What is Process Mining?
• What are key terms in Process Mining?
• What are the benefits of Process Mining?
• Why do companies use Process Mining analytics?

Process Mining combines different analytical fields to analyze operational processes based on their actual "as-is" data, which is available as log files in ERP-Systems. It's derived from the field of Data Mining as it uses similar technology to mine processes for improvements.

The goal of Process Mining is to transform data into actionable insights by revealing the current state of processes and create opportunities for improvement.

In comparison to Data Mining, Process Mining focuses on business processes and its transactional data in ERP systems.

The idea of Process Mining is to discover, monitor, and improve business processes.

Throughout processes and operations, every detail gets recorded. These details leave behind a trace called digital footprints and can be found throughout different systems in an organization. Process Mining captures these footprints and allows you to visualize them in a step-by-step journey view.

The most system based tasks get recorded in log files, so-called event logs.

Process mining uses event logs and process-related data to "mine" processes and better understand them. The full understanding of processes include:

• Identifying process trends, patterns, and deviations
• Detailed visualization of actual processes
• Discovering new ways to increase process efficiency

All executed steps are revealed and can be analyzed for conformance and inefficiencies (such as rework or redundant tasks). Process Mining uses techniques from Business Intelligence to visualize this information and hence provide valuable insights.

Based on the findings, the respective process improvements can be defined. For example:

• Implementing approval steps to ensure compliance

• Removing redundant tasks to reduce cycle time
• Change process flows for efficient execution

When we talk about Process Mining, it's important to clarify the common terms. Let's use an example to explain the key terms. Imagine you work at a company selling T-shirts with custom prints to customers (customers can also buy ones with existing print). The following terms below are commonly used.

Process Model: Most companies already have existing process models, which visually describe the sequence of tasks and responsibilities ("to-be" process). The process model below, visualizes all necessary tasks "to be" executed in order to complete the process.

Process cases: A Case defines a one-time execution of the process (complete or incomplete). The blue dots represent the different department tasks. The red dot represents the tasks that have been executed in the process. 

Process variants: A variant is a set of cases with the same process flow (same choices, activities, events). Only small differences (different end-decision) will lead to a new variant. The example below shows 3 different process variants (based on 4 cases).

Events: In Process Mining, events define every action within a system and these actions have a specific timestamp. Assuming the process is fully executed, each individual task is considered an event since we receive a system-generated timestamp on each task. For example, when the shirt is sent for printing, this is considered an event and has a time-stamp associated with this specific task.

Attributes: Besides the specific tasks and time-stamps, there is other information related to the process called Attributes. Task attributes provide additional details on the executed tasks for certain cases, such as cost center, type of execution (user or system), and costs. Case attributes provide additional details on cases, such as order type, method of payment, and choice of shipping speed. See example below for case attributes.

Large business processes often contain millions of cases, which are distributed in hundreds of different variant. The core of Process Mining is to identify the actual as-is state of processes and reveal the inefficient variants to improve processes.

So far, we've learned that Process Mining captures digital footprints of system-based tasks to deliver insights into the actual operational process. Sounds great! But what exactly are these insights? 

Process Mining has three important capabilities:

• Process Discovery
• Process Conformance
• Process Performance

Process Discovery visualizes the way executed tasks are performed and backtracks the process flow.

This helps companies to:

• Identify the actual "as-is" process in a system
• Identify process deviations, exceptions, or outliers
• Identify the critical path (the longest sequence of tasks from start to finish) and the total number of executions
• Get a starting point to dig deeper into processes

Process Conformance focuses on the mapping of "as-is" data to an existing predefined "to-be" process, e.g. a BPMN process model.

This helps companies to:

• Identify the total of non-conform cases
• Identify the root course of each case
• Develop measures to ensure conformance

Process Performance is about measuring the process and define performance indicators that are shown in a dashboard.

This helps companies to:

• Create and evaluate measures for improvements
• Benchmark performance (e.g. of different regions)
• Monitor the overall performance process in real-time to react more quickly in case of changes

There are several reasons why companies investigate and improve their operational processes, but the main one is a common reason that impacts almost every business.

Digital Transformation is one of the most used terms in today's world and will be a challenge for many companies in the next years.

Digital Transformation means the change in process of a company to adapt to the fast changing global market and customer expectations by using digital technologies and implementing a digital infrastructure. It combines the latest innovative tools and processes with the company's expertise in order to meet customer expectations and ensure a competitive position for the business.

This means that many companies need to adapt their processes in a way which new technologies can be implemented to meet customer expectations and gain new customers. Previous analog processes need to be transformed into digital ones.

Digital transformation combines the latest innovative tools and processes with the company's expertise in order to meet customer expectations and ensure a competitive position for the business.

• Big Data and Real-Time Process Analytics

  Data mining helps identify trends, patterns, and customer behavior in a large data set. Combined with Process Mining, this allows a seamless stream of operational process and production data for real-time analytics.

• Cloud Technology

  Cloud technologies allow for remote resources, storage and process data on demand, and the reduction of hardware and infrastructure costs. Cloud-based Process Mining tools provide powerful connectors to many systems and analyze millions of cases within seconds to deliver process insights.

• Artificial Intelligence (AI) and Machine Learning (ML)

  AI gives companies a completely new basis for decision-making, enabling them to make more informed decisions. Due to its rapid development, it's already implemented in many industries, including IT, financial services, automobile manufacturing and healthcare. AI in Process Mining, is based on actual "as-is" data and can predict the process for cases, for example, incoming orders.

Alright! Now that you've learned the role of Process Mining with digital transformation, let's now look at the impact and benefits of Process Mining.

As you already know, the core focus of Process Mining is to reveal the invisible as-is state of a process. How can companies leverage from this?

The success of a company's business depends on its own process understanding. The data availability of a process is a prerequisite for a data analysis with Process Mining tools. However, once the data has been uploaded to the mining tool, all executed steps are visible, which now also reveals inefficiencies and enables the company to develop steps for improvement.

• Process standardization

  Process Mining identifies variants, outliers and non-conform cases. Also, manual changes and task repetitions (loops) become transparent, which helps companies focus on the root cause and better standardize the process.

• Reduced costs

  Once we know how the process runs, programmers can build RPA bots to automate processes. The transparent "as-is" state of a process also reveals which steps are outdated due to changes and can now be eliminated in order to prevent unnecessary costs.

• Faster reaction time

  Due to the constant live monitoring through dashboards for process KPI, companies can quickly notice if a problem occurs and take the necessary steps to resolve this.

• Resolved bottlenecks

  The time spent analyzing each step becomes less consuming and reveals the steps that are time-consuming and slow down the overall process. Once this is identified, the company can either simplify those steps, allocate more resources to them, or automate them.

• Ensured compliance

  Once the actual steps and variants are revealed, they can easily be compared with an existing "to-be" process (e.g. an underlying process model) to show non-conform tasks (e.g. missing approvals) or uncovered flows through the process. This helps companies identify and implement measures for ensuring process compliance.

Processes can only be mined if they're fully executed in one or several systems. There are common processes that are suitable for processes and often reveal findings worth looking deeper into.

Now that you know the basics, let's take a look at process mining in practice.