Task Scoring
Standard
A clear, fair, consistent way to recognize impact. Every task is evaluated on 4 weighted dimensions, then mapped to one or more performance pillars with contribution weights.
4 Weighted Dimensions
Every task is scored 1–5 on each dimension. The weighted total (out of 100) determines the star rating.
Quality / Correctness
How accurate, reliable, and complete the output is. Meets requirements and holds up to validation.
Impact / Value
The value created for users, stakeholders, or the business. Adoption, decision influence, efficiency, cost.
Timeliness
How well work was delivered against the agreed timeline. On-time, ahead, or behind schedule.
Engineering Maturity
How well work is designed, documented, automated, and enables future use or maintenance.
Scope Change Fairness Rule: If a task is deprioritized, paused, or blocked by external dependencies, evaluation is based on progress made vs. plan, quality of work completed, and clarity of findings/handover — not on final impact or adoption.
Star Rating Thresholds
After scoring all 4 dimensions and computing the weighted total (out of 100), the result maps to a star rating.
3 Star
70 – 79
out of 100
4 Star
80 – 89
out of 100
5 Star
90 – 100
out of 100
Scores 1–2 (below 70) indicate significant gaps requiring targeted improvement.
Scoring Guidelines by Work Type
Select the work type that best describes the task, then apply the rubric to score each dimension.
Production / Delivery
Dashboards, pipelines, models in production, APIs, scheduled jobs
| Score | Quality / Correctness (40%) | Timeliness (20%) | Impact / Value (30%) | Engineering Maturity (10%) |
|---|---|---|---|---|
| 5 | Exceeds requirements. Highly accurate, stable, no rework needed. | Significantly ahead of schedule with buffer. | High adoption (≥90%) and measurable positive business impact. | Introduced automation, reusable components, or scalable solution. |
| 4 | Meets requirements with high quality. Minor improvements only. | Ahead of schedule or optimized delivery time well. | Meaningful value added — improved efficiency, usability, or data quality. | Well-documented, clean and maintainable implementation. |
| 3 | Requirements fully met. Correct and stable output. No major issues. | Delivered on time (within agreed deadline). | Value delivered as expected; useful to intended users. | Basic documentation and maintainable implementation. |
Aggregation Formula & Example
Pillar scores are weighted averages of task scores. Simple averaging would assume equal task relevance — the weighted approach reflects actual contribution.
Pillar Score Formula
Pillar Score (1–5) =
Σ (Task Score × Pillar Weight)
Σ (Pillar Weights)
where weights are each task's % contribution to this specific pillar
Worked Example — Calculating the Delivery Score
| Task | Task Score | Delivery Weight | Score × Weight |
|---|---|---|---|
| ML Pipeline Automation | 4.5 | 40% | 4.5 × 0.40 = 1.80 |
| Incident Root Cause Resolution | 5.0 | 20% | 5.0 × 0.20 = 1.00 |
| Stakeholder Dashboard Project | 4.0 | 50% | 4.0 × 0.50 = 2.00 |
Sum of Weighted Scores
1.80 + 1.00 + 2.00 = 4.80
Sum of Weights
0.40 + 0.20 + 0.50 = 1.10
Delivery Score
4.80 ÷ 1.10 = 4.36 / 5
Why not a simple average? (4.5 + 5.0 + 4.0) ÷ 3 = 4.50. Assumes every task contributed equally to Delivery. The weighted result (4.36) reflects the actual relevance of each task to this specific pillar — more accurate, more fair.
Key Takeaways
- Higher task score + higher pillar weight → greater impact on that pillar's final score.
- Only the portions mapped to a specific pillar affect it — irrelevant task work stays out.
- Strong performance in one pillar does not automatically increase any other.
Task Influence Cap
No single task should contribute more than 25%of a pillar's overall weight allocation. This prevents one exceptional task from masking sustained underperformance across the rest of the period.
If a task maps 50% of its impact to Delivery, but only 2–3 tasks contributed to Delivery total, that task is unusually dominant — flag for calibration review.
Step-by-Step Guide
Classify the Work Type
Identify whether the task is Production/Delivery, POC/Research, or Operational/Support. This determines which rubric applies.
Score Each Dimension 1–5
Using the rubric for your work type, score Quality, Timeliness, Impact/Value, and Engineering Maturity independently.
Multiply by Weight
Quality × 40, Impact × 30, Timeliness × 20, Engineering × 10. Sum those four to get your total out of 100.
Map to Star Rating
70–79 = 3★, 80–89 = 4★, 90–100 = 5★. This task score becomes evidence that feeds into the 7-pillar Performance Evaluation.
Map Task to Performance Pillars
Separately from the 4-dimension weights, assign pillar contribution weights — these answer a different question: not "how good was the task?" but "which performance areas does this task count toward?" Weights must total 100% across all pillars for that task.
Example · Step 5 in action — pillar weight assignment
After scoring ML Pipeline Automation across the 4 dimensions and arriving at a task score of 4.5★, the manager then separately assigns pillar relevance weights based on what the task actually involved:
Delivery
Core production deployment, end-to-end ownership.
Technical Excellence
Automation, scalable architecture, MLOps practices.
Reliability
Pipeline monitoring, data quality gates.
Collaboration
No significant stakeholder coordination required.
Leadership, Development, and Company Culture are not relevant to this task → 0%. Total across all 7 pillars = 100%. The task score of 4.5 now feeds into Delivery, Technical Excellence, and Reliability — each weighted by the % assigned here.
Task scores are evidence. Pillar scores are judgment-backed.
Together, they create fair, balanced, and meaningful performance evaluations.