Calculation Breakdown
Trellis scores are designed to measure developer productivity across various aspects of their work. This topic aims to help you can better interpret Trellis scores and use them to guide improvements in developer productivity.
Constants
Maximum Score: 90 (This is the default maximum score for any metric)
Maximum Reference Multiplication Factor: 10 (This factor is used when a metric value exceeds the threshold)
Threshold: This is defined in the Trellis profile for each metric. It represents either the best performance (100% score) or worst performance (0% score) baseline, depending on the metric.
Weights: These are also defined in the Trellis profile and determine how much each factor (like Quality, Impact, Volume, etc.) contributes to the overall score.
Key terminologies
Metric Score
The Metric Scores represents how well a developer is performing on a particular metric compared to a predefined benchmark or threshold.
Factor Score
A factor score is a composite score that represents performance across a group of related metrics. It provides a broader view of performance in a specific area of development work. Factor scores are weighted based on their importance.
Step 1: Metric Score Calculation
Each metric (e.g., number of PRs per month, percentage of rework) is scored on a scale of 0-90, with the possibility of reaching 99 for exceptional performance.
Metric Scores in Trellis can be categorized in two categories:
Best Performance Metrics: Higher values are better (e.g., Number of PRs per month)
Worst Performance Metrics: Lower values are better (e.g., Percentage of Legacy Rework)
Case A: Best Performance Metrics (100% score)
Using Number of PRs per month as an example:
Threshold (Best Performance) = 20 PRs per month
If Metric Value = 0, Metric Score = 0
If Metric Value <= Threshold:
If Metric Value >= Threshold * Maximum Reference Multiplication Factor then Metric Score = 99
Otherwise: Metric Score = 90 + { [(Metric Value * 100) / (Threshold * Maximum Reference Multiplication Factor)] } * 10 / 100
Case B: Worst Performing Metrics (0% score)
Let's use the Percentage of Rework metric as an example.
Threshold (Worst Performance) = 75% rework
If Metric Value = 0, Metric Score = Maximum Score i.e. 90
Otherwise: 100 - Minimum of [100, (Metric Value * 100 / Threshold)]
These calculations ensure that performance is fairly evaluated across different types of metrics.
Step 2: Factor Score Calculation
Metrics are grouped into factors like Quality, Impact, Volume, etc. To calculate the factor scores system calculates the average of its metric scores. Each factor is assigned a weight based on its importance. This weight is multiplied with the average of its metric scores.
Let's calculate the score for the Quality factor, which includes two metrics:
Percentage of Rework: Score 60 (from previous example)
Percentage of Legacy Rework: Score 70 (assumed for this example)
Assuming the weight for Quality factor as 8 (out of 10)
Step 3: Overall Trellis Score Calculation
The overall Trellis score is the average of all the Factor Scores.
Let's assume we have Factor Scores for all six factors:
Quality: 52 (from previous example)
Impact: 48
Volume: 72
Speed: 63
Proficiency: 81
Leadership & Collaboration: 70
Last updated