Skip to main content

4.2 KPI Dictionary (OEE definitions, loss taxonomy, calculation rules)

4.2 KPI Dictionary (OEE Definitions, Loss Taxonomy, Calculation Rules)

A KPI (Key Performance Indicator) is a navigation instrument, not a "report card." If the dashboard says "Green" but the line is stopped, the dashboard is lying. To manage operations digitally, you must strip away "Vanity Metrics" and enforce strict, mathematical definitions for every number displayed on the floor.

The Gold Standard: OEE (Overall Equipment Effectiveness)

OEE is the rigorous measure of value-added time. It punishes every second the machine is not producing good parts at maximum theoretical speed.

The Formula: OEE = Availability × Performance × Quality

Availability (A)

  • Definition: The percentage of scheduled time the machine was actually running.
  • Formula: Run_Time / Planned_Production_Time
  • The Trap: Do not deduct "unplanned breaks" from the denominator to make the number look better.
  • Rule: If the shift is 8 hours and you have 1 hour of breakdowns, Availability is 87.5%, not 100%.

Performance (P)

  • Definition: The speed at which the machine ran compared to its theoretical maximum.
  • Formula: (Total_Count / Run_Time) / Ideal_Run_Rate
  • The Trap: Never use "Budgeted Rate" or "Average Rate." Use the "Nameplate Rating" (Physics limit).
  • Rule: If the machine can do 100 UPH (Units Per Hour) but you schedule it for 80 UPH, running at 80 results in 80% Performance, not 100%.

Quality (Q)

  • Definition: The percentage of units that were good the first time (First Pass Yield).
  • Formula: (Total_Count - Defect_Count) / Total_Count
  • The Trap: Do not count "Reworked Units" as Good in this metric. Rework hides the cost of failure.

Canonical KPI Dictionary

Adopt these immutable definitions. If Finance calculates "Efficiency" differently, they are calculating a financial variance, not an engineering reality.

Metric

Formula / Logic

Source

Refresh

Owner

OEE

A × P × Q

MES

Real-time

Ops Mgr

TEEP

OEE × (Planned_Time / 24 Hours)

MES

Shiftly

Plant Mgr

MTBF

Total_Run_Time / Count_of_Failures

CMMS/MES

Weekly

Maint Mgr

MTTR

Total_Downtime / Count_of_Failures

CMMS/MES

Weekly

Maint Mgr

FPY

(Input - Fails) / Input (First Pass Only)

MES (Test)

Real-time

Quality

Utilization

Run_Time / 24 Hours

Machine State

Real-time

Planner

Cycle Time

Timestamp_Out - Timestamp_In

MES (Tracking)

Real-time

Process Eng

Loss Taxonomy (The Reason Tree)

You cannot fix "Efficiency Loss." You can only fix "Feeder Jam on Slot 4." Adhere to a strict hierarchical tree for downtime categorization (ISA-95 standard).

Level 1: The Major State

  1. Running: Producing units.
  2. Idle: No demand (Starved/Blocked).
  3. Unplanned Down: Failure/Error.
  4. Planned Down: Maintenance, Setup, Break.

Level 2: The Root Cause (Mandatory)

When a machine stops, the MES must prompt the operator (or read the PLC error code) to assign a Level 2 Reason.

  • Availability Losses (Stops):
    • Equipment Failure: Motor, Sensor, Belt, Software Crash.
    • Setup/Adjust: Changeover, Calibration, Warm-up.
    • Material: Starvation (No Input), Blockage (Buffer Full), Material Empty.
  • Performance Losses (Slowdowns):
    • Micro-stops: < 2 Minutes (Jams, Mis-picks). Auto-categorize these.
    • Speed Loss: Running at reduced rate due to quality risk.
  • Quality Losses (Defects):
    • Scrap: Material destroyed.
    • Rework: Process repeated.

Calculation Rules & Governance

Algorithms must be consistent across all lines.

The "Ideal Cycle Time" Constant

  • Definition: The absolute fastest time the machine can process one unit (e.g., 10.5 seconds).
  • Governance: Stored in the MES Master Data (Product-Resource relation).
  • Lock: Only Process Engineering can update this value.
  • Rule: If Performance > 105%, your Standard Cycle Time is wrong. Fix the Master Data.

Handling Micro-Stops

Operators cannot explain every 30-second stop. It destroys their workflow.

  • Rule: If Downtime < 2 Minutes → Then Auto-tag as "Micro-stop / Minor Stoppage."
  • Analysis: Review the aggregate "Micro-stop" bucket weekly to identify chronic issues.

Starvation vs. Blockage (The Conjoined Twins)

  • Starvation: Machine is ready, but upstream has sent no parts. (Fault lies Upstream).
  • Blockage: Machine is ready, but downstream conveyor is full. (Fault lies Downstream).
  • Logic: The MES must read the sensors at the in-feed and out-feed to auto-classify these states. Do not ask the operator to guess.

Changeover Time (SMED)

  • Start Trigger: Last Good Piece of Old Product.
  • End Trigger: First Good Piece of New Product.
  • Measurement: This includes run-down, setup, material loading, and first-article validation.

Final Checklist

Category

Metric / Control

Threshold / Rule

Math

OEE Availability

Denominator is Planned Production Time, not Total Calendar Time.

Speed

Ideal Rate

Based on "Nameplate" (Physics), not "Budget" (Finance).

Taxonomy

Hierarchy

All downtime maps to standard Tree (Avail/Perf/Qual).

Logic

Micro-stops

Stops < 2 mins are auto-coded. No user input required.

Quality

FPY

Reworked units = 0% Quality for that pass.

Governance

Master Data

Cycle Times locked by Engineering.

Drift

Performance Cap

If Performance > 105% → Flag Master Data Error.