Scoring

The QC Automation Agent uses a two-step weighted scoring system to convert individual check results into a single quality score per operator. This page explains how that scoring works so anyone reviewing the QC board can interpret the numbers.


How Check Results Become Numbers

Each of the 29 checks produces a result. These results map to numeric values:

Result Score Meaning
YES 1.0 Data is present and meets the criteria
PARTIAL 0.5 Data is partially present or partially meets the criteria
NO 0.0 Data is missing or does not meet the criteria
N/A Excluded This check does not apply; it is removed from the calculation
INCONCLUSIVE 0.0 The agent could not determine the answer; scored as zero

Checks marked N/A are excluded entirely – they do not count against the operator. This prevents wells from being penalized for checks that genuinely do not apply to their situation.


Step 1: Category Averages

The 29 checks are grouped into 7 categories. Within each category, the agent calculates the average score of all applicable checks.

Example: If the BHA category has 6 checks and one well’s results are YES, YES, PARTIAL, YES, NO, YES, the category average is:

(1.0 + 1.0 + 0.5 + 1.0 + 0.0 + 1.0) / 6 = 0.75

If any check in the category returns N/A, it is simply excluded from the average. The denominator adjusts accordingly.


Step 2: Weighted Total

Each category carries a weight reflecting its relative importance to overall data quality. The agent multiplies each category average by its weight, sums the results, and divides by the total weight to produce the final score.

Category Weights

Category Weight Rationale
BHA 5 Critical operational data. BHA records are essential for drilling optimization, failure analysis, and regulatory compliance.
Trajectory and Anti-Collision 5 Critical operational data. Survey and well plan data is fundamental to safe directional drilling.
Live Data 4 Real-time monitoring. Active data feeds provide immediate operational visibility.
Drilling Reports 3 Operational records. Daily reporting documents operations and supports post-well analysis.
Engineering 2 Planning documents. Engineering data is typically established before drilling and changes less frequently.
Tool Inventory 2 Asset tracking. Equipment records support logistics and tool management.
File Drive 1 Document storage. Supplementary uploads that back up structured platform data.

Total weight: 22

flowchart LR
    subgraph "Step 1: Category Averages"
        A1["BHA\n6 checks"] --> B1["Avg: 0.75"]
        A2["Trajectory\n5 checks"] --> B2["Avg: 0.90"]
        A3["Live Data\n4 checks"] --> B3["Avg: 1.00"]
        A4["Reports\nup to 5 checks"] --> B4["Avg: 0.60"]
        A5["Engineering\n3 checks"] --> B5["Avg: 0.67"]
        A6["Tool Inv.\n2 checks"] --> B6["Avg: 0.50"]
        A7["File Drive\n4 checks"] --> B7["Avg: 1.00"]
    end

    subgraph "Step 2: Weighted Total"
        B1 -- "x5" --> C["Weighted\nSum"]
        B2 -- "x5" --> C
        B3 -- "x4" --> C
        B4 -- "x3" --> C
        B5 -- "x2" --> C
        B6 -- "x2" --> C
        B7 -- "x1" --> C
        C --> D["Final Score\n0.78"]
    end

    style D fill:#4a90d9,stroke:#2c5aa0,color:#fff

Worked Example

Consider a fictitious operator, “Example Energy,” with one well called “EXAMPLE 1H.” Here are the check results and how they produce a final score:

Individual Check Results

Category Check Result Score
BHA BHA Distribution YES 1.0
BHA BHA Comments PARTIAL 0.5
BHA BHA Uploads YES 1.0
BHA BHA Failure Reports YES 1.0
BHA BHA Component Completeness YES 1.0
BHA Post-Run BHA Grading NO 0.0
Trajectory Surveys YES 1.0
Trajectory Survey Program YES 1.0
Trajectory Survey Corrections YES 1.0
Trajectory EDM Files YES 1.0
Trajectory Well Plans N/A
Live Data WITSML Connected YES 1.0
Live Data Live Geosteering YES 1.0
Live Data NPT Tracking YES 1.0
Live Data Cost Analysis YES 1.0
Reports Mud Report Distro YES 1.0
Reports Mud Program N/A
Reports Formation Tops NO 0.0
Reports Drilling Program YES 1.0
Reports AFE Curves NO 0.0
Engineering Roadmaps N/A
Engineering Wellbore Diagrams YES 1.0
Engineering Engineering Scenarios NO 0.0
Tool Inventory Rig Inventory Data YES 1.0
Tool Inventory Tool Catalog Data NO 0.0
File Drive File Drive: BHAs YES 1.0
File Drive File Drive: Well Plans YES 1.0
File Drive File Drive: Drill Prog N/A
File Drive File Drive: Mud Reports YES 1.0

Category Averages

Category Calculation Average
BHA (1.0 + 0.5 + 1.0 + 1.0 + 1.0 + 0.0) / 6 0.75
Trajectory (1.0 + 1.0 + 1.0 + 1.0) / 4 1.00
Live Data (1.0 + 1.0 + 1.0 + 1.0) / 4 1.00
Reports (1.0 + 0.0 + 1.0 + 0.0) / 4 0.50
Engineering (1.0 + 0.0) / 2 0.50
Tool Inventory (1.0 + 0.0) / 2 0.50
File Drive (1.0 + 1.0 + 1.0) / 3 1.00

Note: N/A results are excluded, so the denominator adjusts. Trajectory has 4 checks counted instead of 5 because Well Plans returned N/A.

Weighted Total

Category Average Weight Contribution
BHA 0.75 5 3.75
Trajectory 1.00 5 5.00
Live Data 1.00 4 4.00
Reports 0.50 3 1.50
Engineering 0.50 2 1.00
Tool Inventory 0.50 2 1.00
File Drive 1.00 1 1.00
Total   22 17.25

Final Score: 17.25 / 22 = 0.784 (78.4%)


Interpreting Scores

Scores range from 0.0 (no data present) to 1.0 (all applicable checks passed). As a general guide:

  • 0.90 and above – Excellent. Nearly all data modules are complete and current. Minor gaps only.
  • 0.70 to 0.89 – Good, with room for improvement. Most critical data is present, but some modules need attention.
  • 0.50 to 0.69 – Needs attention. Significant gaps in data completeness, likely across multiple categories.
  • Below 0.50 – Critical gaps. Major data modules are incomplete, affecting operational visibility.

These are general interpretations, not hard pass/fail thresholds. The appropriate standard may vary by operator, basin, or well phase.


Reading the QC Board

The Monday.com QC board displays one row per operator with the following information:

  • Agent Score – The overall weighted score (the number calculated above)
  • Per-check status columns – Individual result for each of the 29 checks across all wells for that operator
  • Metadata – Operator name, number of wells, last run timestamp

The agent only updates columns where the score has changed since the last run, making it easy to see which operators’ data quality has improved or declined.