Overview
The Variant Performance calculator analyzes performance metrics across different process variants, helping you identify which process paths are most efficient and where bottlenecks occur. Unlike Variant DNA which focuses on frequency and distribution, Variant Performance provides detailed duration statistics including mean, median, maximum, and total duration for each variant. This enables data-driven decisions about which process paths to optimize, which to promote as best practices, and where to allocate improvement resources.
The calculator supports flexible configuration including minimum case count thresholds to ensure statistical significance, sorting by multiple metrics, and limiting results to top performers.
Common Uses
- Identify the fastest and slowest process variants to prioritize optimization efforts
- Compare variant performance to establish best practice pathways
- Monitor variant duration trends to detect process degradation over time
- Allocate resources based on variant complexity and performance characteristics
- Benchmark different process paths to guide standardization initiatives
- Detect performance outliers within variants using maximum duration analysis
Settings
Maximum Number of Variants: Limits results to the top N variants based on your sorting criteria. Default is 10. Setting this to a higher number (e.g., 20) provides more comprehensive analysis, while lower numbers (e.g., 5) focus on the most relevant variants.
Sort By: Determines which performance metric to use for ranking variants. Options include:
- Case Count: Sorts by the number of cases following each variant (useful for finding high-volume paths)
- Mean Duration: Sorts by average case duration (default, best for identifying typical performance)
- Median Duration: Sorts by middle duration value (less affected by outliers)
- Maximum Duration: Sorts by longest case in each variant (highlights worst-case scenarios)
- Total Duration: Sorts by cumulative duration across all cases (shows overall capacity consumption)
Sort Order: Controls whether highest or lowest values appear first:
- Descending: Shows variants with highest values first (e.g., slowest average duration)
- Ascending: Shows variants with lowest values first (e.g., fastest average duration)
Duration Unit: Specifies the time unit for all duration metrics. Options include Seconds, Minutes, Hours, Days (default), Weeks, Months, and Years. Choose based on your process timeframe - use Hours for operational processes, Days for standard business processes, or Weeks/Months for long-running processes.
Minimum Case Count: Filters out variants with fewer cases than this threshold. Default is 10. This ensures statistical significance by excluding rare variants with insufficient data. Set to 0 to include all variants regardless of frequency.
Examples
Example 1: Optimizing Order Fulfillment Performance
Scenario: Your order fulfillment process has multiple variants, and you want to identify which paths deliver the fastest turnaround while maintaining sufficient volume to be statistically meaningful.
Settings:
- Maximum Number of Variants: 10
- Sort By: Mean Duration
- Sort Order: Ascending
- Duration Unit: Days
- Minimum Case Count: 20
Output:
The calculator reveals performance across your top 10 variants:
| Variant | Cases | Mean | Median | Max | Total |
|---|---|---|---|---|---|
| Variant 3 | 245 | 2.1 days | 2.0 days | 4.5 days | 514.5 days |
| Variant 1 | 892 | 3.2 days | 3.0 days | 8.1 days | 2,854.4 days |
| Variant 5 | 156 | 4.8 days | 4.2 days | 12.3 days | 748.8 days |
| Variant 2 | 634 | 8.5 days | 7.9 days | 18.2 days | 5,389.0 days |
Variant 3 shows the fastest mean duration at 2.1 days with 245 cases - enough volume to be statistically significant. Variant 1, your highest-volume path with 892 cases, averages 3.2 days - only slightly slower but processing far more cases. Variant 2 shows significantly longer duration at 8.5 days average.
Insights: Variant 3 represents your best practice path. Analyzing what makes it efficient (direct shipping, pre-verified inventory, streamlined approvals) can help optimize other variants. Variant 1's high volume and reasonable performance make it a good baseline. Variant 2's poor performance (8.5 days average) combined with high volume (634 cases) represents the largest improvement opportunity - investigate root causes like backorder delays or approval bottlenecks.
Example 2: Identifying Performance Outliers in Invoice Processing
Scenario: You want to identify process variants with the worst-case performance issues by examining maximum durations, which reveal cases that take exceptionally long regardless of average performance.
Settings:
- Maximum Number of Variants: 15
- Sort By: Maximum Duration
- Sort Order: Descending
- Duration Unit: Days
- Minimum Case Count: 10
Output:
Sorting by maximum duration reveals problem variants:
| Variant | Cases | Mean | Median | Max | Total |
|---|---|---|---|---|---|
| Variant 8 | 47 | 12.5 days | 9.2 days | 45.3 days | 587.5 days |
| Variant 12 | 23 | 8.1 days | 6.8 days | 38.7 days | 186.3 days |
| Variant 4 | 156 | 6.2 days | 5.5 days | 32.1 days | 967.2 days |
| Variant 1 | 892 | 5.2 days | 4.8 days | 18.5 days | 4,638.4 days |
Variant 8, despite only 47 cases, shows a maximum duration of 45.3 days - more than double any other variant. The large gap between median (9.2 days) and maximum (45.3 days) indicates severe outlier cases. Variant 12 also shows concerning maximum duration (38.7 days) relative to its median (6.8 days).
Insights: The extreme maximum durations in Variants 8 and 12 suggest specific cases getting stuck in exception handling, approval escalations, or rework loops. Click on these variants to drill down into the longest-running cases and identify common characteristics (specific suppliers, high amounts, missing documentation). These outliers likely represent process failures that need exception-handling improvements rather than average process optimization.
Example 3: Capacity Planning Using Total Duration Analysis
Scenario: You need to understand which process variants consume the most overall capacity to allocate resources effectively and prioritize automation efforts.
Settings:
- Maximum Number of Variants: 20
- Sort By: Total Duration
- Sort Order: Descending
- Duration Unit: Hours
- Minimum Case Count: 5
Output:
Sorting by total duration reveals capacity consumption:
| Variant | Cases | Mean | Median | Max | Total |
|---|---|---|---|---|---|
| Variant 1 | 2,340 | 18.5 hrs | 16.2 hrs | 45.3 hrs | 43,290 hrs |
| Variant 2 | 1,567 | 24.3 hrs | 21.8 hrs | 68.4 hrs | 38,078 hrs |
| Variant 4 | 892 | 36.7 hrs | 32.1 hrs | 89.2 hrs | 32,736 hrs |
| Variant 7 | 456 | 48.2 hrs | 44.5 hrs | 112.3 hrs | 21,979 hrs |
Variant 1 consumes 43,290 total hours despite having the fastest mean duration (18.5 hours) due to its high volume (2,340 cases). Variant 4 shows slower performance (36.7 hours mean) but still ranks third in total capacity consumption. Variant 7 has the slowest mean (48.2 hours) but fewer cases.
Insights: Total duration analysis reveals where process capacity is actually being consumed. Variant 1's 43,290 hours represents the largest improvement opportunity - even small per-case reductions (e.g., cutting 2 hours from the 18.5 hour average) yield massive capacity savings (4,680 hours annually). This makes Variant 1 the prime candidate for automation despite already having good mean performance. Variant 4 offers medium-term opportunity - its slower performance and significant volume warrant investigation.
Example 4: Monitoring Process Standardization Progress
Scenario: After implementing process standardization, you want to track whether performance is converging across variants by examining median durations, which are less affected by outliers than mean values.
Settings:
- Maximum Number of Variants: 10
- Sort By: Median Duration
- Sort Order: Ascending
- Duration Unit: Days
- Minimum Case Count: 15
Output:
Before standardization:
| Variant | Cases | Mean | Median | Max |
|---|---|---|---|---|
| Variant 1 | 445 | 4.2 days | 3.8 days | 15.2 days |
| Variant 3 | 234 | 6.5 days | 5.9 days | 18.7 days |
| Variant 5 | 156 | 9.8 days | 8.2 days | 24.5 days |
| Variant 8 | 89 | 14.3 days | 12.1 days | 38.9 days |
After standardization:
| Variant | Cases | Mean | Median | Max |
|---|---|---|---|---|
| Variant 1 | 823 | 3.9 days | 3.7 days | 11.2 days |
| Variant 2 | 612 | 4.2 days | 4.0 days | 12.5 days |
| Variant 3 | 387 | 4.8 days | 4.5 days | 13.8 days |
The median duration range narrowed from 3.8-12.1 days (8.3 day spread) to 3.7-4.5 days (0.8 day spread), demonstrating successful standardization. More cases now follow the top 3 variants (1,822 vs 835 previously), and maximum durations dropped across the board.
Insights: Process standardization successfully reduced performance variability. The tight clustering of median durations (3.7-4.5 days) indicates consistent execution across variants. The reduction in maximum durations shows better exception handling. The shift in case distribution toward top variants demonstrates adoption of standardized paths. This validates the standardization initiative and provides a baseline for ongoing monitoring.
Example 5: Comparing High-Volume vs Low-Volume Variant Performance
Scenario: You want to understand if high-volume variants perform differently than low-volume variants by adjusting the minimum case count threshold and analyzing results.
Settings:
- Maximum Number of Variants: 10
- Sort By: Case Count
- Sort Order: Descending
- Duration Unit: Days
- Minimum Case Count: 100
Output:
High-volume variants (min 100 cases):
| Variant | Cases | Mean | Median | Max | Total |
|---|---|---|---|---|---|
| Variant 1 | 1,245 | 5.2 days | 4.8 days | 18.5 days | 6,474 days |
| Variant 2 | 892 | 6.3 days | 5.9 days | 22.3 days | 5,620 days |
| Variant 3 | 634 | 7.1 days | 6.5 days | 28.7 days | 4,501 days |
| Variant 4 | 456 | 4.8 days | 4.5 days | 15.2 days | 2,189 days |
Now change Minimum Case Count to 10 and re-run:
| Variant | Cases | Mean | Median | Max |
|---|---|---|---|---|
| Variant 27 | 23 | 2.1 days | 2.0 days | 4.5 days |
| Variant 15 | 45 | 2.8 days | 2.6 days | 6.8 days |
| Variant 4 | 456 | 4.8 days | 4.5 days | 15.2 days |
| Variant 1 | 1,245 | 5.2 days | 4.8 days | 18.5 days |
Low-volume variants (Variants 27 and 15) show significantly faster performance (2.1-2.8 days mean) compared to high-volume variants (4.8-7.1 days mean).
Insights: The performance disparity reveals that low-volume variants likely represent expedited or simplified paths (rush orders, simple products, VIP customers) while high-volume variants handle standard, more complex cases. Before attempting to scale the fast low-volume processes, verify they can handle the characteristics of high-volume cases. Alternatively, segment cases to route simple cases through expedited variants, potentially reducing overall processing time by directing appropriate cases to faster paths.
Output
The calculator produces a performance table with the following metrics for each variant:
Variant: Identifier for the process variant (e.g., Variant 1, Variant 2). Click on the variant name to drill down and see the complete activity sequence and all cases following this path.
Case Count: Number of cases following this variant. Higher case counts provide more statistical confidence in the performance metrics.
Mean Duration: Average duration across all cases in the variant, expressed in your selected time unit. Best for understanding typical performance when outliers are not a major concern.
Median Duration: Middle duration value when cases are sorted by duration. More robust than mean when outliers are present, representing the typical case performance.
Maximum Duration: Longest case duration within the variant. Useful for identifying worst-case scenarios, SLA compliance issues, and exception handling problems.
Total Duration: Sum of all case durations in the variant. Reveals overall capacity consumption and helps prioritize improvement efforts based on total organizational impact.
Interactive Features:
- Click any variant row to filter the process view to cases in that variant
- Sort by clicking column headers to reorder results dynamically
- Export data for further analysis in spreadsheet or BI tools
- Combine with filters to analyze variant performance for specific time periods or case segments
Visualization Options: The table can be used in dashboards alongside:
- Variant DNA calculator to see frequency and performance together
- Process maps filtered by variant to visualize different paths
- Time-based charts to track variant performance trends over time
This documentation is part of the mindzie Studio process mining platform.