case duration category for activity

Overview

The Case Duration Category for Activity enrichment analyzes the total time spent on a specific activity across all its occurrences within each case, then categorizes this duration into performance levels: Fast, Normal, Slow, or Extreme. This enrichment is particularly valuable for identifying cases where certain activities are consuming excessive time, enabling focused process improvement efforts.

Unlike simple duration calculations, this enrichment automatically determines optimal performance thresholds based on your actual data distribution using statistical percentiles. It sums up all durations for the selected activity within each case, providing a comprehensive view of how much time is being invested in that particular activity throughout the case lifecycle. This makes it ideal for analyzing activities that may occur multiple times within a case, such as approvals, reviews, or quality checks.

Common Uses

Purchase Order Processing: Categorize cases based on total approval time to identify orders stuck in lengthy approval cycles
Claims Processing: Analyze total investigation time across multiple investigation activities to spot complex claims requiring excessive review
Manufacturing Quality Control: Measure cumulative inspection time to identify products requiring repeated quality checks
Customer Service: Track total customer interaction time across multiple support contacts to identify high-maintenance cases
Loan Applications: Categorize applications by total underwriting time to streamline the approval process
Healthcare Patient Flow: Analyze total diagnostic testing time to optimize patient pathways
Software Development: Measure total code review time across multiple review cycles to improve development velocity

Settings

Activity Name: Select the activity you want to analyze for performance categorization. The enrichment will sum up all occurrences of this activity within each case and categorize the total duration. This is required and must match an existing activity name in your event log exactly.

Case Filter (Optional): Apply filters to limit which cases are analyzed. You can filter by case attributes, time ranges, or other criteria. Only filtered cases will be used to calculate the performance thresholds, and only these cases will receive the new performance category attribute. Cases outside the filter will not be categorized.

Fast Duration Threshold: The maximum duration (in hours) for a case to be categorized as "Fast". If left at default (0), the system automatically calculates this as the 20th percentile of all case durations for the selected activity. For example, if set to 2 hours, any case with total activity duration up to 2 hours is considered Fast.

Normal Duration Threshold: The maximum duration (in hours) for a case to be categorized as "Normal". If left at default (0), the system automatically calculates this as the 80th percentile. Cases with durations between the Fast threshold and this value are considered Normal performance.

Slow Duration Threshold: The maximum duration (in hours) for a case to be categorized as "Slow". If left at default (0), the system automatically calculates this as the 90th percentile. Cases between the Normal and Slow thresholds are considered Slow, while anything beyond this threshold is categorized as "Extreme".

Examples

Example 1: Invoice Approval Performance

Scenario: A finance team wants to identify invoices that spend excessive time in approval processes, as these often indicate issues requiring manual intervention or complex validation.

Settings:

Activity Name: "Approve Invoice"
Case Filter: None (analyze all invoices)
Fast Duration Threshold: 0 (auto-calculate)
Normal Duration Threshold: 0 (auto-calculate)
Slow Duration Threshold: 0 (auto-calculate)

Output: The enrichment creates a new case attribute called "Approve Invoice - Case Performance" with values:

Fast: Invoices with total approval time in the bottom 20% (e.g., under 2 hours)
Normal: Invoices with typical approval times (e.g., 2-24 hours)
Slow: Invoices taking longer than usual (e.g., 24-72 hours)
Extreme: Invoices with exceptional delays (e.g., over 72 hours)

Insights: The finance team can now filter for "Extreme" cases to investigate why certain invoices require excessive approval time, potentially revealing missing documentation, disputed amounts, or process bottlenecks.

Example 2: Patient Diagnostic Testing

Scenario: A hospital wants to categorize emergency department visits based on the total time spent in diagnostic testing (X-rays, CT scans, lab work) to improve patient flow and resource allocation.

Settings:

Activity Name: "Perform Diagnostic Test"
Case Filter: Department = "Emergency"
Fast Duration Threshold: 1 (hour)
Normal Duration Threshold: 3 (hours)
Slow Duration Threshold: 5 (hours)

Output: New attribute "Perform Diagnostic Test - Case Performance" categorizes each patient visit:

Fast: Patients with minimal testing (under 1 hour total)
Normal: Standard diagnostic workup (1-3 hours)
Slow: Complex cases requiring extensive testing (3-5 hours)
Extreme: Critical cases with prolonged diagnostics (over 5 hours)

Insights: The hospital identifies that "Extreme" cases often involve multiple specialists and could benefit from a dedicated complex case coordinator to reduce waiting times between tests.

Example 3: Software Code Review Cycles

Scenario: A development team wants to understand which pull requests are consuming excessive review time across multiple review iterations, impacting deployment velocity.

Settings:

Activity Name: "Code Review"
Case Filter: Repository = "Core Platform"
Fast Duration Threshold: 0 (auto-calculate)
Normal Duration Threshold: 0 (auto-calculate)
Slow Duration Threshold: 0 (auto-calculate)

Output: Creates "Code Review - Case Performance" attribute with automatically calculated thresholds:

Fast: PRs reviewed quickly (e.g., under 4 hours total across all reviews)
Normal: Standard review time (e.g., 4-16 hours)
Slow: PRs requiring extensive review (e.g., 16-24 hours)
Extreme: PRs with exceptional review effort (e.g., over 24 hours)

Insights: Analysis reveals that "Extreme" cases typically involve architectural changes or lack proper documentation, leading the team to implement better PR templates and architectural review processes.

Example 4: Manufacturing Quality Inspections

Scenario: A manufacturing plant needs to identify products requiring excessive quality inspection time, which often indicates production issues or design flaws.

Settings:

Activity Name: "Quality Inspection"
Case Filter: Product Line = "Premium Series"
Fast Duration Threshold: 0.5 (30 minutes)
Normal Duration Threshold: 2 (hours)
Slow Duration Threshold: 4 (hours)

Output: New attribute "Quality Inspection - Case Performance" with values:

Fast: Products passing quickly (under 30 minutes total inspection)
Normal: Standard inspection time (30 minutes - 2 hours)
Slow: Products requiring detailed inspection (2-4 hours)
Extreme: Products with major quality issues (over 4 hours)

Insights: Products in the "Extreme" category correlate strongly with specific production batches, revealing equipment calibration issues that can be proactively addressed.

Output

This enrichment creates a new case attribute with the following characteristics:

Attribute Name: "[Activity Name] - Case Performance" (e.g., "Approve Invoice - Case Performance")

Data Type: String (categorical)

Possible Values:

Fast: Total activity duration falls within the fastest performing cases (below Fast threshold)
Normal: Total activity duration is within typical range (between Fast and Normal thresholds)
Slow: Total activity duration is longer than normal (between Normal and Slow thresholds)
Extreme: Total activity duration exceeds expected boundaries (above Slow threshold)
Negative: Rare cases where calculated duration is negative (data quality issue)
Null: Cases where the selected activity doesn't occur

How Thresholds Work: When thresholds are set to 0 (default), the system automatically calculates them using:

Fast Threshold: 20th percentile of all positive durations
Normal Threshold: 80th percentile of all positive durations
Slow Threshold: 90th percentile of all positive durations

This statistical approach ensures categories are meaningful for your specific process, automatically adapting to your data's natural distribution.

Using the Output: The new performance category attribute can be used in:

Performance Dashboards: Visualize the distribution of cases across performance categories
Root Cause Analysis: Filter for "Extreme" cases to investigate process issues
Process Comparison: Compare performance categories across different case attributes
Predictive Analytics: Use as a feature for predicting case outcomes or delays
SLA Monitoring: Track percentage of cases in each category against targets

Overview

Common Uses

Settings

Examples

Example 1: Invoice Approval Performance

Example 2: Patient Diagnostic Testing

Example 3: Software Code Review Cycles

Example 4: Manufacturing Quality Inspections

Output

See Also