Categorize Duration For Activity

Overview

The Categorize Duration for Activity enrichment transforms activity durations into performance categories that provide immediate insights into process efficiency. This enrichment analyzes how long each instance of a selected activity takes to complete and automatically assigns performance labels such as "Fast", "Normal", "Slow", or "Extreme" based on statistical thresholds. Instead of working with raw duration values that can be difficult to interpret at scale, you get clear performance indicators that instantly highlight which activities are meeting expectations and which require attention.

This enrichment is particularly powerful for performance monitoring and optimization initiatives. It uses intelligent percentile-based calculations to automatically determine appropriate thresholds based on your actual data distribution, ensuring that categories reflect the real performance patterns in your process. The enrichment handles edge cases such as negative durations (which can occur when timestamps are incorrect) by categorizing them separately, and it creates a new event attribute that can be used in performance dashboards, filters, and detailed process analysis to identify bottlenecks and improvement opportunities.

Common Uses

• Monitor approval activity performance to identify which approvals are taking longer than expected
• Track manufacturing step durations to ensure production processes meet cycle time targets
• Analyze customer service response times to identify agents or cases needing performance improvement
• Measure payment processing speeds to detect delays in financial workflows
• Evaluate medical procedure durations to optimize hospital resource allocation
• Monitor order fulfillment activities to identify warehouse or shipping bottlenecks
• Assess document review times in compliance processes to prevent regulatory delays

Settings

Activity Name: Select the specific activity whose duration you want to categorize. This dropdown lists all activities present in your event log. The enrichment will analyze the duration of each occurrence of this activity (the time between its start and end timestamps) and assign a performance category to each instance.

Fast Duration Threshold: The maximum duration (in hours, minutes, and seconds) that will be categorized as "Fast" performance. Any activity instance completing in this time or less will be labeled as "Fast". If left at the default value of 00:00:00, the system will automatically calculate this threshold using the 20th percentile of all positive durations for the selected activity.

Normal Duration Threshold: The maximum duration that will be categorized as "Normal" performance. Activity instances with durations greater than the Fast threshold but less than or equal to this value will be labeled as "Normal". If left at 00:00:00, the system automatically uses the 80th percentile of durations.

Slow Duration Threshold: The maximum duration that will be categorized as "Slow" performance. Activity instances exceeding the Normal threshold but within this limit will be labeled as "Slow". If left at 00:00:00, the system automatically uses the 90th percentile. Any duration exceeding this threshold will be categorized as "Extreme".

Filter: Optional filter to apply before calculating categories. This allows you to focus the categorization on specific subsets of your data, such as particular time periods, regions, or case types. The filter affects both the automatic threshold calculation and which events receive category labels.

Examples

Example 1: Purchase Order Approval Performance

Scenario: A procurement team needs to monitor the performance of their purchase order approval process. They want to identify which approvals are being processed efficiently and which are experiencing delays.

Settings:

• Activity Name: "Approve Purchase Order"
• Fast Duration Threshold: 00:00:00 (auto-calculate)
• Normal Duration Threshold: 00:00:00 (auto-calculate)
• Slow Duration Threshold: 00:00:00 (auto-calculate)
• Filter: None

Output: Creates an event attribute called "Approve Purchase Order - Performance" with values:

• "Fast" for approvals completed in under 2 hours (20th percentile)
• "Normal" for approvals taking 2-8 hours (20th to 80th percentile)
• "Slow" for approvals taking 8-24 hours (80th to 90th percentile)
• "Extreme" for approvals exceeding 24 hours (above 90th percentile)
• "Negative" for any data quality issues with negative durations

Insights: The procurement team can now easily identify that 20% of approvals are processed very quickly, while 10% are taking an extremely long time. They can drill down into the "Extreme" cases to understand what causes these delays and implement targeted improvements.

Example 2: Patient Triage Classification

Scenario: A hospital emergency department wants to categorize triage assessment times to ensure patients are being evaluated within appropriate timeframes based on medical best practices.

Settings:

• Activity Name: "Triage Assessment"
• Fast Duration Threshold: 00:05:00 (5 minutes)
• Normal Duration Threshold: 00:10:00 (10 minutes)
• Slow Duration Threshold: 00:15:00 (15 minutes)
• Filter: Emergency_Department = "Main ED"

Output: Creates "Triage Assessment - Performance" attribute where:

• Assessments under 5 minutes are "Fast"
• Assessments from 5-10 minutes are "Normal"
• Assessments from 10-15 minutes are "Slow"
• Assessments over 15 minutes are "Extreme"

Insights: Hospital administrators can monitor compliance with triage time targets, identify peak periods where assessments slow down, and ensure critical patients receive timely evaluation.

Example 3: Manufacturing Quality Inspection

Scenario: A manufacturing plant needs to monitor quality inspection durations across different production lines to maintain consistent throughput while ensuring thorough inspections.

Settings:

• Activity Name: "Quality Inspection"
• Fast Duration Threshold: 00:00:00 (auto-calculate)
• Normal Duration Threshold: 00:00:00 (auto-calculate)
• Slow Duration Threshold: 00:00:00 (auto-calculate)
• Filter: Product_Category = "Electronics"

Output: The enrichment analyzes only electronics inspections and creates performance categories:

• "Fast" for inspections under 12 minutes
• "Normal" for inspections from 12-35 minutes
• "Slow" for inspections from 35-45 minutes
• "Extreme" for inspections over 45 minutes

Insights: Production managers can balance inspection thoroughness with production speed, identify inspectors who may need additional training, and detect potential quality issues when inspection times are consistently extreme.

Example 4: Loan Application Processing

Scenario: A bank wants to categorize loan application processing times to meet service level agreements and improve customer satisfaction.

Settings:

• Activity Name: "Process Loan Application"
• Fast Duration Threshold: 24:00:00 (1 day)
• Normal Duration Threshold: 72:00:00 (3 days)
• Slow Duration Threshold: 120:00:00 (5 days)
• Filter: Loan_Type = "Personal Loan"

Output: Creates performance categories for personal loan processing:

• Applications processed within 1 day are "Fast"
• Applications taking 1-3 days are "Normal"
• Applications taking 3-5 days are "Slow"
• Applications exceeding 5 days are "Extreme"

Insights: The bank can track SLA compliance, identify process improvements for slow applications, and potentially offer expedited processing as a premium service for "Fast" category achievements.

Example 5: Customer Support Ticket Resolution

Scenario: A software company wants to analyze the performance of their technical support team's ticket resolution activities across different priority levels.

Settings:

• Activity Name: "Resolve Technical Issue"
• Fast Duration Threshold: 00:00:00 (auto-calculate)
• Normal Duration Threshold: 00:00:00 (auto-calculate)
• Slow Duration Threshold: 00:00:00 (auto-calculate)
• Filter: Priority = "High"

Output: For high-priority tickets, the enrichment creates categories:

• "Fast" for resolutions under 30 minutes (top 20% performers)
• "Normal" for resolutions from 30 minutes to 2 hours
• "Slow" for resolutions from 2-4 hours
• "Extreme" for resolutions exceeding 4 hours

Insights: Support managers can identify which high-priority issues are resolved quickly, recognize top-performing agents, and investigate extreme cases to improve training and documentation.

Output

When executed, this enrichment creates a new event attribute named "[Activity Name] - Performance" that contains performance category values for each occurrence of the selected activity. The attribute characteristics include:

• Attribute Type: Event attribute (attached to individual activity instances)
• Data Type: String (text)
• Possible Values:
  • "Fast" - Duration is at or below the Fast threshold
  • "Normal" - Duration exceeds Fast but is at or below Normal threshold
  • "Slow" - Duration exceeds Normal but is at or below Slow threshold
  • "Extreme" - Duration exceeds the Slow threshold
  • "Negative" - Duration is less than zero (indicates data quality issues)
  • Null - Activity has no duration information available

The performance attribute integrates seamlessly with other mindzieStudio features:

• Use in performance dashboards to visualize activity efficiency distribution
• Apply filters to focus on extreme or slow activities for process improvement
• Combine with conformance checking to correlate performance with process violations
• Export to business intelligence tools for detailed performance analytics
• Use in calculators to compute average performance across different dimensions

See Also

• Duration Between Two Activities - Calculate time between different activities instead of categorizing single activity performance
• Categorize Attribute Values - Create custom categories for any numerical attribute beyond just durations
• Durations Between Case Attribute and Activity Times - Measure time from case start to specific activities
• Performance Filter Category - Apply performance categorization across multiple attributes simultaneously

---

This documentation is part of the mindzie Studio process mining platform.