Simulation Results Explanation: Uncover the Why Behind Your Process Delays

A Deeper Look into Queues, Resources, and Delays

Simulations in IBM Process Mining allow organizations to explore “what-if” scenarios, for instance, what would happen if automation were introduced, staffing levels changed, or certain process steps were optimized.They help estimate key outcomes such as processing time reduction, cost savings, or Return on Investment (ROI) before any real implementation takes place.

While simulations already offered powerful insights, one question often remained unanswered: 
Where exactly are the delays coming from, and which resources or activities are causing them?
The Simulation Results Explanation feature was designed to answer that , bringing visibility, transparency, and interpretability to simulated results.

Why Simulation Explanations Matter

Previously, users could view aggregate simulation results such as throughput times or bottleneck activities. However, they couldn’t see when queues formed, which roles were overloaded, or why certain activities were delayed.

Now, IBM Process Mining goes beyond surface metrics , it infers queue and resource information from the simulated event log, making it visible through a clear, interactive dashboard.

This new capability helps users:

• Identify the root causes of process delays

• Detect resource bottlenecks and imbalance

• Understand activity-wise queue behavior and timing

• Support data-driven staffing and optimization decisions
  

Exploring the Simulation Results Explanation Dashboard

The dashboard provides multiple layers of analysis, from high-level time trends to role-specific drilldowns allowing users to interpret simulation results in a stepwise, intuitive way.

Queue Time and Activity Trends Over Time

• This chart visualizes:

  • Average Queue Time - shown as a blue trend line

  • Total Delayed Activities (Queues) - represented as green bars

• Each point on the chart represents a time slice during the simulation, helping users identify when queues were most severe and how many activities were affected.

• How users can analyze:

  • Peaks in the line graph signal periods of congestion, such as month-end closures or approval surges.

  • If queue times rise while delayed activities remain low, it may indicate prolonged delays in fewer activities (perhaps due to scarce specialized resources).

  • Conversely, spikes in delayed activities with moderate queue time suggest short queues but higher frequency, possibly reflecting workflow imbalance.

  • By selecting a date directly from this chart, users can filter all subsequent drilldown results to that specific period, focusing analysis where it matters most.

   Value to users:

  This chart helps pinpoint when bottlenecks occur and whether they are systemic or time-dependent, guiding capacity planning or scheduling decisions.

Activity Performance Metrics

• This visualization compares five performance measures across activities:

  • Queue Time - Additional waiting time (adding up to the expected waiting time) occurring for an activity during the simulation, because of no resources available to take the activity in charge.

  • Expected Wait Time -The minimum waiting time occurring for an activity, as specified in the Simulation's evolutions (or derived from the baseline's event log).

  • Working Time - The time a resource or robot is allocated to complete an activity, as specified in the Simulation's evolutions (or derived from the baseline's cost settings). The higher this time is, the more likely queues will occur during the Simulation."

  • Activity Frequency - Number of times the activity was executed

  • Delayed Activities (Queues) - Delayed executions of an activity because of missing workforce availability. Each delayed activity corresponds to a queue time.

  Each activity is represented on a bar chart, with additional line plots overlaying execution frequency and queue occurrence trends.

• How users can analyze:

  • Activities with high queue time and low working time indicate inefficient resource allocation, they wait longer than they’re being worked on.

  • High working time with minimal queue suggests well-balanced roles or activities that are continuously worked upon without idling.

  • If an activity has frequent occurrence and high delayed activity count, it’s a strong candidate for automation or capacity increase.

  • Comparing expected wait time vs. queue time helps distinguish between predictable vs. unexpected delays, valuable for SLA forecasting.

  Value to users:

  This analysis enables users to prioritize process improvements at the activity level, identifying where queues form most often and whether they stem from high demand, slow execution, or resource unavailability.
  

  
  

Role Performance Metrics

• Here, performance is aggregated by role, offering a clear view of how resource allocation influences queues.

  It displays:

  • Total Queue Time 

  • Total Delayed Activities (count)

• How to analyze:

  • Roles like DIRECTOR with high queue time and many delayed activities show workforce overload.

  • Roles like BACK-OFFICE with low queue time may have excess or balanced capacity.

  • Cross-referencing with activity metrics helps pinpoint which tasks performed by that role cause bottlenecks.

   What users gain:

  This insight allows managers to optimize staffing levels, plan role reassignments, or introduce automation to roles under heavy workload.
  It offers a tangible, data-backed approach to improving operational balance and efficiency.

Summary Widget and Drill-down Analysis
   

Summary Widget

Lists the top activities with the longest queue times, alongside:

• The number of delayed activities

• The roles responsible

• Average queue durations

This quick summary provides an instant view of which areas deserve immediate attention.

Drill-down Results

Users can interact with charts or use dropdown filters for:

• Time Range

• Activity

• Role

The Drill-down Results panel shows the top 10 queue events by default, detailing:

• Queue duration

• Associated activity and role

• Timestamp of occurrence

For deeper analysis, users can export the complete dataset via “Download CSV”, which includes every event with a queue, enabling offline analytics or AI-based process diagnostics.

Insights Beyond the Charts

The true power of the Simulation Results Explanation feature lies in its analytical depth.
It doesn’t just visualize data, it helps you reason with it.

Some example insights users can derive include:

• Identifying time windows of systemic congestion (e.g., quarter-end workloads)

• Recognizing understaffed roles leading to recurring queues

• Validating whether automation in a specific task can reduce delays effectively

• Comparing simulated “to-be” results with the baseline process to measure real improvement potential

By combining these perspectives, users can form a complete cause-and-effect view of process behavior, enabling proactive decisions instead of reactive fixes.

From Simulation to Action

The Simulation Results Explanation feature transforms IBM Process Mining from a simulation platform into an insight-driven improvement tool.
It gives business users, analysts, and process owners the transparency they need to:

• Understand why certain outcomes appear in simulation

• Validate where optimization will yield the most benefit

• Make informed choices on resource allocation, automation, and process redesign

By revealing the hidden dynamics of queues and resource constraints, this enhancement ensures that every simulation becomes a blueprint for smarter, faster, and more resilient operations.