Plan vs Actual Output Variance

Goal of the analysis:

Quantify and explain the gap between planned production output and actual output by line, shift, SKU, and plant. The analysis isolates actionable drivers—mix, yield/quality, schedule adherence, changeover overruns, speed loss, downtime (breakdowns, starved/blocked), and material/labor constraints—and translates variance into capacity time at the constraint and service impact (OTIF/backlog). Executives use it to stabilize delivery, reduce expediting/overtime, prioritize TPM/SMED and material readiness, calibrate planning realism (finite capacity vs infinite), and decide on debottlenecking or capex only where justified.

Data required:

• Planning baselines:
  • Master Production Schedule (MPS), finite capacity schedules/dispatch lists by line and shift, weekly build plan by SKU, and frozen window rules.
  • AOP/S&OP volumes and weekly buckets for context; planned changeovers and standard durations (changeover matrix).
• Production execution:
  • Work order actuals: SKU, planned vs actual start/finish, planned vs produced good quantity, scrap/rework, hold/release timestamps.
  • Run/stop events from MES/SCADA (breakdowns, starved/blocked, micro-stops), actual changeover times, speed/rate data.
• Standards and capacity:
  • Standard cycle times or standard hours per unit by SKU/operation; nameplate rates; crew sizes; bottleneck resource identification.
  • OEE (Availability, Performance, Quality) or components by line/shift to convert hours to effective capacity.
• Materials and inventory readiness:
  • ATP/CTP status; kit completeness, component shortages/substitutions, supplier delivery confirmations and delays.
  • WIP/FG inventory positions and quality holds affecting release.
• Labor and maintenance context:
  • Timeclock hours by line/shift (regular/overtime, agency), absenteeism; maintenance PM calendar, MTBF/MTTR, breakdown logs.
• Service and demand linkage:
  • Customer orders/backlog by SKU/date, OTIF, expedite requests; Book-to-Bill to contextualize backlog build/burn.
• Normalization:
  • Seasonality, NPIs and learning curves, engineering changes affecting standard times; policy on planned stops; time zone/calendar alignment.

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

1. Align scope and definitions.
   • Choose the horizon (daily/shift/weekly) and unit (good units or standard hours). Define “plan” baseline: MPS at freeze for near term, weekly build plan for beyond.
   • Define attainment window (e.g., within the frozen 48–72 hours) and include/exclude approved reschedules consistently.
2. Extract and reconcile plan vs actual.
   • Join planned orders to actual execution by order/line/shift. Reconcile planned quantities to dispatch list and actuals to MES/ERP; align timestamps.
   • Validate that totals tie to period production and inventory movements; flag missing/duplicate orders and anomalous rates.
3. Compute headline variance and attainment.
   • Output Variance (units) = Actual good units − Planned units; Attainment (%) = Actual ÷ Plan × 100.
   • Convert to standard hours at the bottleneck: Earned Hours (Actual) vs Planned Standard Hours to remove mix effects.
   • Translate variance into capacity days at the constraint = Variance standard hours ÷ Effective daily capacity.
4. Decompose drivers (build bridges).
   • Create a variance waterfall for each line/plant:
     • Plan → Mix effect (plan vs actual SKU mix at planned volumes using standard hours) → Schedule adherence/timing (orders not run in window) → Changeover overrun (actual − planned setup time) → Downtime (breakdowns, starved/blocked, PM overruns) → Speed loss (Performance vs standard) → Yield/quality (scrap/rework) → Material shortages (idle time) → Labor shortfall/absence → Other holds → Actual.
   • Use minutes-based loss bridges: Available run minutes → lost to setup/downtime/speed → earned minutes → good units.
5. Segment and localize issues.
   • Segment variance by line/work center, shift/crew, SKU family, and customer priority. Rank by impact (units/$) and repeatability.
   • Overlay with OEE and schedule adherence to distinguish true capacity constraints from readiness/planning flaws.
6. Link to service and backlog.
   • Map output shortfall to at-risk orders by date/customer; compute OTIF impact and backlog days added at the constraint.
   • Quantify expedites/overtime triggered by misses.
7. Trend and stability.
   • Plot weekly attainment and variance components for 12–16 weeks; annotate SMED/TPM events, supplier issues, NPIs. Compute variability (StdDev) by line/shift.
8. Scenario and what-if.
   • Simulate SMED (−30–50% setup), speed centerlining (+5–10% Performance), targeted PM, kitting to 98% kit completeness, added split shift; estimate variance recovery and backlog clearance date.
   • Test plan realism: finite schedule with current OEE vs infinite; quantify required OEE or shifts to meet plan.
9. Integrity checks.
   • Ensure consistent inclusion of planned stops across plan and actual; avoid double-counting material waits as both downtime and changeover overrun.
   • Confirm standard times reflect current engineering; validate “good units” align with quality release and scrap logs.

Format of the output of analysis:

• Executive scorecard: Plan vs Actual (units and standard hours), Attainment %, variance in capacity days at constraint, OTIF/backlog impact, top 5 variance drivers.
• Waterfall bridges: variance decomposition (mix, schedule adherence, setup overrun, downtime, speed, yield, materials, labor) by line and plant.
• Heatmaps: attainment by line × shift and by SKU family; reason-code minutes by line.
• Gantt variance charts: planned vs actual order start/finish, swaps, and overruns.
• Loss bridges: minutes from available to earned to good output; Pareto of downtime and setup overruns.
• Scenario panel: expected attainment lift and backlog reduction from SMED/TPM/kitting/centerlining/overtime/extra shift.

How to interpret results:

• Negative variance dominated by setup overruns and sequence swaps: Planning discipline and SMED opportunity; batch by family, enforce frozen window, externalize setup tasks.
• Negative variance with high downtime (breakdowns/starved): Reliability and material readiness issues; strengthen TPM/condition monitoring; fix kitting and supplier SLAs.
• Low Performance (speed loss) with normal Availability: Centerlining and micro-stop elimination required; review sensors, guides, lubrication, and operator standard work.
• Good units shortfall driven by quality: FPY issue; tighten process windows, first-article checks, and supplier/material specs.
• Attainment misses only on the bottleneck: Throughput constrained—focus improvement there before increasing labor on non-constraints.
• Plan frequently exceeded (Actual > Plan): Possible conservative planning or underutilized capacity; consider raising plan, rebalancing mix, or reducing overtime if used to overachieve.
• Variance volatile week-to-week: Unstable process or plan realism issues; standardize definitions and adopt finite capacity scheduling.

Steps a company can take to improve on this measure:

• Planning and scheduling:
  • Adopt finite capacity scheduling using demonstrated OEE; set frozen windows (24–72h) and clear expedite governance; batch by family to reduce changeovers.
  • Improve standards (cycle times, changeover matrix) and reflect engineering changes promptly.
• Execution and reliability:
  • Run SMED on top recipe pairs; externalize tasks, quick-change tooling; resequence to minimize swaps.
  • TPM and condition monitoring on constraint assets; align PMs to demand valleys; improve spares and MTTR.
  • Centerline speeds/parameters; eliminate micro-stops; enforce Andon responses.
• Material readiness and supply:
  • Implement kitting/line-side supermarkets; shortage dashboards; supplier ASNs and delivery SLAs; qualify alternates for fragile components.
  • Use ATP/CTP to commit dates aligned to material and capacity realities.
• Labor and staffing:
  • Cross-train crews on the constraint; stabilize staffing; manage overtime to avoid productivity decay; add split shifts temporarily to recover backlog.
• Governance and linkage to service:
  • Daily tiered reviews at the line with plan vs actual boards; track minutes-based loss Pareto and owners; tie actions to OTIF/backlog recovery.
  • Integrate with S&OP; adjust lead-time quotes/pricing during recovery to prevent backlog spikes.
• Scenario guidance:
  • If weekly attainment is 82% with 25% of minutes in setup, run SMED on top 3 family swaps (target −40% setup) and enforce 48h freeze; expect +8–12 pts attainment in 6–8 weeks.
  • If downtime drives 60% of loss (starved + breakdowns), launch kitting (≥98% kit completeness) and install vibration monitoring on the filler; target −50% starved time and +6–8 pts attainment.
  • If speed loss is 10 pts, execute centerlining and preventive tasks (sensors/guides); expect +3–5 pts attainment and lower scrap.

Benchmark comparisons:

General benchmarks (directional):

• Frozen-window plan attainment: World-class 90–95%+ per line/shift; typical 80–90% in make-to-stock, 70–85% in high-mix make-to-order.
• Variance guardrails: Weekly output variance within ±5–10% at plant level in stable environments; bottleneck lines monitored to tighter bands (±5%).
• Setup adherence: Actual vs planned setup time within ±10–20% for mature SMED programs.
• Speed/Performance: Running rate ≥90–95% of standard for centerlined lines; micro-stops <5% of run time.

Constructing internal benchmarks:

• Track 12–16 weeks of attainment and variance components by line/shift/SKU family; publish quartiles and set alerts (e.g., attainment <85% for 3 consecutive weeks triggers action).
• Maintain demonstrated capacity and standard times; recalibrate plan realism quarterly; compare finite plan accuracy vs actuals.
• Pair plan variance with OTIF/backlog and OEE; prioritize lines with high variance and high service impact.
• After major SMED/TPM or portfolio changes, reset baselines and tighten guardrails progressively.