Don’t Roll the Dice with Process Mining Simulation

Introduction

Would you roll the dice to make business decisions? Most likely not. Business decisions are expected to be based on, among other criteria, thorough analysis of timely data, holistic problem-solving, and consideration of the current and future state of the business and its environment. However, when assessing, prioritising, and implementing process improvements, many companies ignore these principles and rely on back-of-the-envelope maths or the simulation tools that come with their process mapping and mining tools. These Level 1 simulation features have a limited understanding of how improvements affect process KPIs and allow very basic testing of process improvements by changing selected input variables in very restricted ways.

In this blog post, we will demonstrate that using the simulation features that come with mapping and mining software - not to speak of back-of-the-envelope maths - is like rolling the dice to make your business decisions. Process mining simulation may ignore the characteristics of good business decisions, leading to suboptimal decisions that hurt your business’s bottom line and that decision-makers cannot trust or rely on. Moreover, those basic simulation features often do not come with valuable features, including extensible simulation models, process variant analysis, and advanced AI and analytical tools.

Instead, decision-makers should use a simulation tool like Silico that enables simulation at a higher level. Silico mirrors processes accurately, produces reliable insights that decision-makers can trust, and allows decision-makers to change input values over time and observe outcomes as they develop over the future days, weeks, months, and years. Silico then builds on these reliable simulations by, for example, expanding simulations beyond single processes to commercial outcomes, testing new process variants, and automatically identifying optimisations to unlock realms of value.

Process Mapping and Mining may include Limited Simulation Features

Many organisations leverage process mining, process mapping, and business intelligence tools to understand their processes comprehensively. With this knowledge, they can easily unlock the value of simulation to improve their processes by answering the critical question: ‘What’s next?’.

However, it’s important to note that not all simulation features are created equal. When process mapping and mining tools include simulation, they are often limited (Level 1 in the simulation hierarchy). These simulations are easy to use and can quickly convert mining graphs and analytics into simulation models.

However, they have severe limitations. Decision-makers take significant risks when they base their decisions on them due to the inaccurate transformations of mining outcomes to a simulation model, the limits of event logs, and the ignoring of critical variables. Using them is like asking decision-makers to rely on rolling dice to make transformation decisions. As a result, it’s crucial for companies to be aware of the limitations of process mapping and mining simulations and to choose more advanced simulation tools to drive effective process improvements.

5 Critical Shortcomings of Process Mining Simulation

The limitations of many integrated process simulation features become evident when we evaluate them against characteristics of good business decisions:

1. The limited simulations facilitated by process mining tools do not allow for a comprehensive analysis of the process. They only cover the system-based activities included in event logs and cannot account for activities executed outside of recorded systems - leading to an incomplete analysis that underestimates processing times, lead times, and costs while overestimating outcomes like revenue.

2. Simulations provided as part of process mining tools may rely on outdated information and include versions of the process that are no longer used. Process mining graphs use event logs collected over a period of time. If the process changes during data collection, mining graphs and resulting simulation models include the old and altered process.

3. Basic simulation features in process mining software may not capture relevant activities accurately, especially activities that take place in parallel, leading to incorrect impact assessments of transformations. This limitation reduces the accuracy and reliability of simulated KPIs.

4. Often, the limited functionality of process mining-based simulations does not consider the current state of the business, such as backlogs, queues, or work-in-process or (WIP) when starting the simulation, which can significantly impact process and commercial KPIs. This limitation leads to inaccurate simulation outcomes and can misinform decision-makers.

5. Simulation tools coming with process mining software often lack the ability to consider changes over time, reducing their value for generating accurate and valuable insights for decision-makers. Therefore, their simulations may not reflect real-world scenarios, including growing case volumes or process changes happening at specific points in time or over a period of time.

Why you Should not Trust a Basic Process Mining Simulation

The shortcomings of simulation features integrated into process mapping and mining software have significant implications for decision-makers. Decision-makers require accurate and realistic simulations to assess the impact of process changes and prioritise their implementation. The graph below illustrates the impact on forecasting accuracy of, for example, a lead-to-order or order-to-cash simulation ignoring the business’s current state such as orders in the pipeline - the fourth limitation introduced above. Here, just ignoring initial backlogs or queues when starting the simulation leads to significant errors and low accuracy of the simulation. In reality, all five limitations will misguide the estimation of transformation impacts and prioritisation of changes.

These queues in the process are an essential intermediary outcome that affects many process and commercial KPIs, including lead time, required FTEs, revenue, cash flow, and FTE utilisation - leading decision-makers to roll the dice when assessing and prioritising process improvements. Below are some KPI differences between the Silico simulation we tested above and its counterpart that, like many simulation features that come with process mapping and mining software, ignores current business conditions (here: initial WIP).

An adaptable, transparent, and replicable simulation like Silico is required to avoid such inaccuracies. Silico can consider your business’s current conditions and track KPIs accurately - among many other advantages. Therefore, it can generate more reliable simulation outcomes, enabling decision-makers to make informed and data-driven decisions. Advanced simulation tools like Silico then build on these reliable simulations to allow, for example, experimentation with process variants and continuous optimisations.

Don’t Roll the Dice with an Integrated Process Mining Simulation Feature

Generate insights your decision-makers can trust with Silico in just 4 weeks.

Simulating your process changes is valuable, it can reduce the risk of costly errors and helps avoid investing time, effort, and budget into low-impact changes. However, picking the right simulation tool is equally important. Trusting the simulation features that come with many mapping and mining tools can be like rolling the dice. 

Use Silico to quantify the impact of process improvements and prioritise your changes with a tool that your decision-makers can trust. In just four weeks, Silico builds on your process mining insights, develops a Digital Twin of your process, and delivers you the insights required to optimise it. With Silico, you are using a reliable tool that enables a myriad of valuable simulations.