Overall Equipment Effectiveness

Goal of the analysis:

Quantify how effectively production assets convert scheduled time into good output at the right speed and quality. Overall Equipment Effectiveness (OEE) decomposes losses into Availability, Performance, and Quality to expose the “Six Big Losses” (breakdowns, setup/adjustments, idling/minor stops, reduced speed, defects, startup losses). Executives use OEE to increase throughput without capex, stabilize lead times, lower unit costs, and prioritize maintenance, staffing, and digital investments at true bottlenecks. A rigorous OEE program standardizes definitions, integrates data from MES/SCADA/quality systems, segments by product/shift, and links loss Pareto to targeted improvement actions (TPM, SMED, SPC, predictive maintenance).

Data required:

• Time and production context:
  • Shift calendars: scheduled time, planned stops (breaks, meetings, sanitation, no-demand windows).
  • Standard/ideal cycle time by SKU and machine; rated nameplate speed and allowable tolerances.
  • Line configuration and bottleneck machine identification.
• Machine and event signals:
  • Run/stop states and timestamps from MES/SCADA/PLC (state codes for fault, setup, changeover, starved/blocked, micro-stops).
  • Downtime logs with reason codes; mean time between failures (MTBF), mean time to repair (MTTR).
• Output and quality:
  • Total units produced, good units, scrap/rework counts by SKU/lot; first pass yield; startup/warm-up scrap.
  • Quality defect categories (dimensions, contamination, packaging, labeling) and inspection timestamps.
• Changeover and setup:
  • Changeover start/stop timestamps, internal vs external steps (SMED classification), cleaning/sterilization durations, recipe uploads.
• Support systems (for root cause):
  • Maintenance work orders (preventive/corrective), parts availability, condition monitoring sensors.
  • Supply/logistics events (material shortages, downstream backups), operator rosters/training levels.
• Normalization and reference:
  • Product mix (SKU runs, lot sizes), environmental factors (temp/humidity), M&A/new line ramp flags.

Detailed step-by-step instruction on how to conduct the analysis:

1. Define scope and standardize formulas.
   • Planned Production Time = Scheduled Shift Time − Planned Stops (breaks, meetings, sanitation if policy excludes).
   • Run Time = Planned Production Time − Unplanned Stop Time (breakdowns, setups/adjustments, starved/blocked, micro-stops).
   • Availability = Run Time ÷ Planned Production Time.
   • Performance = (Ideal Cycle Time × Total Units Produced) ÷ Run Time (cap at 100%).
   • Quality = Good Units ÷ Total Units Produced.
   • OEE = Availability × Performance × Quality.
   • Optional: OOE (Overall Operations Effectiveness) = OEE × (Planned Production Time ÷ All Calendar Time); TEEP = OEE × (Loading, i.e., Scheduled Time ÷ All Time).
2. Assemble and cleanse data.
   • Extract event logs, counters, and quality data from MES/SCADA/LIMS/ERP; align to a common time base and machine/line identifiers.
   • Map downtime codes to the Six Big Losses; define micro-stop threshold (e.g., <5 minutes) and ensure consistent capture.
   • Validate ideal cycle times by SKU via engineered standards or recent best-day performance; lock a reference table.
3. Compute OEE by line/machine/shift and SKU.
   • For each period (hour/shift/day), calculate Availability, Performance, Quality, and OEE; separate startup windows if policy counts separately.
   • Reconcile counts to physical inventory and scrap logs; flag anomalies (Performance >100%, negative times).
4. Build loss trees and Pareto charts.
   • Availability loss = Planned Production Time − Run Time; subdivide into breakdown, setup/changeover, starved/blocked, micro-stops.
   • Performance loss = Run Time − (Ideal Cycle Time × Total Units); split into speed loss vs small stops if captured.
   • Quality loss = Total Units − Good Units; split into startup scrap vs steady-state defects by category.
5. Segment and contextualize.
   • Segment by product/SKU, lot size, shift, operator team, and maintenance status (pre/post PM).
   • Identify the system bottleneck (lowest OEE or highest time-in-constraint) and quantify flow impact at the line level.
6. Trend and benchmark.
   • Plot 12–16 weeks of OEE and components with annotations (new PM routines, SMED events, training, recipe updates).
   • Compare cells (lines/plants) using consistent definitions; compute internal quartiles and target bands.
7. Root-cause deep dives.
   • For top losses, run 5-Why/Fishbone with data overlays (sensor trends, SPC charts); correlate failures with condition metrics and environmental factors.
   • For setups, apply SMED: classify internal vs external tasks, videotape changeovers, and quantify quick-change opportunities.
8. Action planning and verification.
   • Translate Pareto items into projects (TPM pillars, predictive maintenance, parameter control limits, recipe management, operator training); set expected bps OEE lift and owners.
   • Pilot interventions on the bottleneck; measure pre/post Availability/Performance/Quality to validate impact.
9. Integrity checks.
   • Ensure planned stops are consistently excluded; avoid double-counting micro-stops in both Availability and Performance.
   • Audit cycle-time table and counter integrity; perform periodic manual time studies to validate system data.

Format of the output of analysis:

• Executive scorecard: OEE and its components (Availability, Performance, Quality) by line/machine and shift; bottleneck identification; trend vs target.
• Pareto charts: downtime categories, speed losses, and defect categories (Six Big Losses) with $/throughput impact.
• Bridges: OEE change driver waterfall (setup reduction, breakdowns, speed, quality) period-over-period.
• Heatmaps: OEE by SKU × line; shift/operator cell performance; changeover durations by recipe pair.
• Action tracker: initiatives, owners, expected OEE uplift, milestones, realized benefits (throughput units, hours freed).

How to interpret results:

• OEE ≥85% with balanced components: World-class for discrete/high-speed packaging; focus on fine-tuning (predictive maintenance, parameter control) and mix-driven optimization.
• OEE 60–80% with Availability drag: Breakdowns and long changeovers dominate; prioritize PM/TPM, SMED, spares, and schedule smoothing.
• OEE 60–80% with Performance drag: Chronic speed loss/micro-stops; address starve/block causes, sensors/jams, operator standard work, and material quality.
• OEE constrained by Quality: Scrap/rework or startup losses; tighten process windows (SPC), calibrations, material specs; enhance first-article checks and poka‑yoke.
• High line OEE but plant throughput flat: Non-bottleneck gains; realign focus to the constraint machine/step and flow (Theory of Constraints).
• Volatile OEE: Unstable processes or data definitions; standardize coding and stabilize with PM and parameter controls.

Steps a company can take to improve on this measure:

• Availability (eliminate downtime):
  • Implement TPM (autonomous maintenance, focused improvement); deploy condition monitoring/predictive analytics on critical assets.
  • SMED for changeovers: externalize tasks, quick-release fixtures, preset tools/recipes; schedule sequencing to minimize recipe swaps.
  • Strengthen spares strategy and maintenance planning; reduce MTTR with standard work and visual management.
• Performance (run at ideal speed):
  • Root-cause micro-stops (material jams, sensors, lubrication); improve starvation/blocking via WIP buffers and balanced upstream/downstream rates.
  • Optimize control loops and setpoints; enforce centerlines and parameter audits; train operators on loss identification.
• Quality (make good units first time):
  • Deploy SPC and real-time alarms; calibrate measurement systems (MSA); tighten supplier specs and incoming quality.
  • Poka‑yoke checks, first-article approval, automated vision systems; DMAIC on top defect modes.
• Scheduling and product mix:
  • Sequence to reduce changeovers; right-size lot sizes; move low-volume SKUs off the constraint or to flexible cells.
• Digital and governance:
  • Standardize OEE definitions and downtime codes; deploy real-time dashboards/Andon; daily tiered reviews with Pareto-driven actions.
  • Link incentives to bottleneck OEE and first-pass yield, not just aggregate hours.
• Scenario guidance:
  • If OEE = 62% with 18% Availability loss from setups, run a SMED kaizen on top two recipe pairs; target −40% changeover time and +8–10 pts OEE in six weeks.
  • If Performance is 85% due to micro-stops, implement centerlining and sensor maintenance; aim +5 pts Performance and +3–4 pts OEE.
  • If Quality at startup is 94%, add first-article vision checks and parameter locks; reduce startup scrap by 50% and lift Quality to 98–99%.

Benchmark comparisons:

General benchmarks (directional):

• Discrete manufacturing/assembly: typical OEE 55–75%; top quartile 75–85%.
• High-speed FMCG/packaging: typical 60–80%; best-in-class 80–90% on mature lines.
• Process/continuous (chemicals, paper): OEE often 70–90% with fewer changeovers; quality losses may dominate.
• Component targets: Availability 85–95%, Performance 90–98%, Quality 98–99.9% depending on industry and regulatory stringency.

Constructing internal benchmarks:

• Track OEE by line/shift/SKU for 12–16 weeks; set plant-specific bands and target the bottom quartile cells.
• Normalize by product mix and demand (exclude sanctioned idle time consistently); maintain a single master for ideal cycle times.
• Publish Six Big Losses Pareto monthly; tie capex approval to demonstrated OEE constraints at the bottleneck.