Production Capacity Constraint & Bottleneck Analyzer

Formula

System Capacity = MIN(Station Capacities); Throughput = Bottleneck Capacity × Utilization × Efficiency

System production capacity equals the minimum capacity across all workstations—the bottleneck. Example: Stations with capacities [100, 80, 120, 90] have system capacity of 80 (the minimum). Actual throughput accounts for utilization (how much of capacity is used) and efficiency (actual vs. theoretical output). Effective capacity = Bottleneck capacity × (1 - downtime %) × efficiency %. Example: 80 units/hr bottleneck, 10% downtime, 85% efficiency = 80 × 0.9 × 0.85 = 61.2 units/hr practical capacity. The formula works because production is serial process—all stations must complete for finished product. The slowest station limits overall pace (chain is only as strong as weakest link). Improving any non-bottleneck station from 100 to 150 changes system capacity by 0 because bottleneck (80) still limits. Only improving bottleneck (80 to 100) increases system capacity to next-slowest station (90), yielding new capacity of 90. This focuses optimization: ONLY invest in bottleneck until it's no longer the constraint.

Worked Examples

Example 1: Manufacturing Line Capacity Analysis

Problem:4-station line: Cutting 100/hr, Assembly 80/hr, QC 120/hr, Packing 90/hr. Current load 85/hr. Target 100/hr. Can we meet target?

Solution:Capacity Analysis: - Cutting: 100/hr capacity, 85/hr load (85% utilization) - Assembly: 80/hr capacity, 85/hr load (106% - IMPOSSIBLE) → Actual load limited to 80/hr - QC: 120/hr capacity, 85/hr load (71%) - Packing: 90/hr capacity, 85/hr load (94%) Bottleneck: Assembly (80/hr) - Lowest capacity - Limits entire system to 80/hr Current Output: 80/hr (system capacity) Target: 100/hr Gap: 20/hr (25% short) Can meet target? NO - Assembly constrains to 80/hr - Must increase Assembly capacity by 20/hr (+25%) Options: Option 1: Upgrade Assembly - Add equipment or worker - Increase 80 → 100/hr - Cost: $50K (new station) - New bottleneck: Packing (90/hr) - System capacity: 90/hr (still short 10/hr) - Need to also upgrade Packing 90 → 100 - Additional cost: $30K - Total: $80K for 100/hr c

Result:System capacity: 80/hr (Assembly bottleneck) | Target: 100/hr (25% short) | Need $25K to reach 90/hr via Assembly + Packing optimization

Frequently Asked Questions

What is a production bottleneck?

Bottleneck is the slowest step in production process limiting overall output. Example: Assembly line has Cutting (100/hr), Assembly (80/hr), Packing (120/hr). Assembly is bottleneck (80/hr). System can only produce 80/hr regardless of other stations' capacity. Theory of Constraints (Goldratt): System output = bottleneck capacity. Improving non-bottleneck stations doesn't increase output. Only improving bottleneck does.

How do I identify bottleneck?

Methods: (1) Observation (queue builds up before bottleneck station), (2) Utilization (station running at 95-100% while others are 60-70%), (3) Cycle time (longest processing time), (4) WIP (work-in-progress accumulates before bottleneck). Example: Parts pile up before Assembly (queue = bottleneck). Assembly workers never idle (high utilization). Other stations have downtime. Assembly is bottleneck. Measure capacity at each stage; lowest capacity = bottleneck.

Should I optimize bottleneck or non-bottleneck?

Always optimize bottleneck first. Theory of Constraints: Hour lost at bottleneck = hour lost for entire system. Hour saved at non-bottleneck = meaningless (just creates more idle time). Example: Assembly (bottleneck) produces 80/hr. Cutting produces 100/hr. Improving Cutting to 120/hr changes nothing (Assembly still limits to 80). Improving Assembly to 90/hr increases system output to 90/hr (+12.5%). Focus: Bottleneck utilization (keep it running), non-bottleneck subordination (support bottleneck).

What is drum-buffer-rope in TOC?

Drum-Buffer-Rope (DBR) is production scheduling method from Theory of Constraints. Drum: Bottleneck sets pace (schedule everything around it). Buffer: Time or inventory before bottleneck (ensures it never starves). Rope: Release raw materials at bottleneck pace (prevents overproduction). Example: Assembly does 80/hr (drum). Maintain 2-hour buffer of parts before Assembly (buffer = 160 units). Release materials to Cutting at 80/hr (rope), not 100/hr (prevents excess WIP).

What happens when I eliminate bottleneck?

Eliminating bottleneck reveals next bottleneck. Example: Assembly 80/hr (bottleneck). Upgrade to 100/hr. Now Packing (90/hr) becomes bottleneck. System capacity: 80 → 90 (+12.5%). Continuous improvement: Identify bottleneck → Improve → New bottleneck emerges → Repeat. Never-ending. But diminishing returns eventually: improving capacity from 80 to 200 may not be valuable if demand is only 150. Match capacity to demand, not infinite optimization.

How do I calculate production capacity?

Capacity = Bottleneck capacity (units/time). Example: Assembly line with 5 stations: A (100/hr), B (80/hr), C (90/hr), D (120/hr), E (85/hr). Bottleneck: B (80/hr). System capacity: 80/hr. Daily capacity (8 hours): 640 units. Accounting for: Downtime (10% breaks, maintenance), efficiency (85% actual vs. theoretical) → Effective capacity: 640 × 0.9 × 0.85 = 490 units/day. Use effective capacity for planning (not theoretical).

What is the cost of bottleneck downtime?

Bottleneck downtime stops entire system. Example: Assembly (bottleneck 80/hr) down for 1 hour. Lost output: 80 units. At $50 profit/unit = $4,000/hour. Non-bottleneck downtime: Only affects that station. Cutting down 1 hour = 0 lost output (Assembly still runs from WIP buffer). Prioritize: Preventive maintenance on bottleneck (minimize downtime), backup for bottleneck (redundancy), less maintenance on non-bottleneck (tolerate occasional downtime).

Should I balance production line (equal capacity)?

Counterintuitive answer: No. Perfectly balanced line (all stations 100/hr capacity) is suboptimal. Why: (1) Variability exists (stations don't actually run at 100% always), (2) Downtime happens (maintenance, breaks), (3) Balanced line has no buffer (any disruption propagates). Better: Bottleneck at 90%, non-bottlenecks at 100-110% (buffer protects bottleneck). Ensure bottleneck never starves. Unbalanced line with deliberate bottleneck and buffers is more robust than balanced line.

How does batch size affect throughput?

Large batches: Fewer changeovers (less downtime) but more WIP (tied-up capital, longer lead times). Small batches: Flexible, responsive, less WIP, but more setups. Optimal batch size depends on: Setup time (if 2 hours setup for 4 hours run, prefer large batches), demand variability (custom orders need small batches), WIP cost (high-value products minimize batches). At bottleneck: Maximize batch size (minimize setup). Non-bottleneck: Flexible (has spare capacity for setup).

What is OEE (Overall Equipment Effectiveness)?

OEE = Availability × Performance × Quality. Availability: Uptime % (planned run time - downtime). Performance: Actual speed vs. theoretical. Quality: Good units / total units. Example: Machine runs 7 hours of 8 (87.5% availability), at 90% of rated speed (90% performance), produces 5% defects (95% quality). OEE = 0.875 × 0.9 × 0.95 = 74.8%. World-class: >85%. Measure OEE at bottleneck—where improvement matters most.

Background & Theory

Production capacity bottleneck analysis identifies constraining workstations limiting overall system throughput using Theory of Constraints principles to prioritize improvement investments and maximize output without wasteful optimization of non-constraining resources. ## Concept Overview Production system capacity equals bottleneck capacity, not sum of station capacities. Example: 5-station line with capacities 100, 80, 120, 90, 110 units/hr. System capacity = 80/hr (bottleneck station). Improving station 3 from 120 to 150 (+30/hr) changes system capacity by 0. Improving station 2 from 80 to 100 (+20/hr) increases system capacity to 90/hr (next bottleneck station 4). This is Theory of Constraints: Focus on constraint exclusively. The financial implication: $100K investment improving bottleneck from 80 to 100/hr yields 25% output increase. Same $100K improving non-bottleneck from 120 to 140/hr yields 0% output increase (wasteful). ROI is infinite on bottleneck investment (within demand limits), zero on non-bottleneck. This focuses capital: spend only where it increases throughput. Operational principle: Non-bottleneck stations should have spare capacity (buffer). This protects bottleneck from starvation. If Cutting (upstream) breaks down but has buffer before Assembly (bottleneck), Assembly continues running. But if Assembly breaks down, entire system stops. Priority: Bottleneck preventive maintenance, backup equipment, focus. ## Key Variables & Intuition • **Station Capacity** — Max throughput (units/hour); theoretical limit • **Current Load** — Actual throughput; utilization = load / capacity • **Bottleneck** — Station with lowest capacity; system constraint • **System Capacity** — Overall throughput; equals bottleneck capacity • **Utilization %** — How much capacity is used; bottleneck should be near 100%, others lower • **Target Output** — Desired throughput; achievable only if ≤ system capacity ## Assumptions • Capacities are accurately measured (theoretical may differ from practical) • Stations are in sequence (not parallel paths) • Bottleneck is stable (doesn't shift hourly based on product mix) • Demand exists for additional output (no point increasing capacity beyond demand) • Quality is consistent (defects don't create rework bottleneck) ## Limitations & Edge Cases • **Multi-product lines** — Bottleneck may shift by product (Product A bottlenecks at Assembly, Product B at Packing) • **Batch production** — Setup times create effective bottlenecks (machine down during changeover) • **Shared resources** — Station serves multiple lines (becomes distributed bottleneck) • **Variable processing time** — Capacity is average; variability creates effective capacity reduction • **Quality issues** — High defect rate at downstream station creates rework bottleneck **Scenario:** Factory identifies Painting (60/hr) as bottleneck, system capacity 60/hr. Invests $200K to upgrade Painting to 100/hr. Celebrates 67% capacity increase. Three months later, output is still 60/hr. Investigation reveals: Drying (downstream) is 60/hr. Painting improvement just moved bottleneck to Drying. Should have analyzed entire line before investing. Now must spend another $150K to upgrade Drying to 100/hr. Lesson: Map entire value stream, identify all constraints, budget for chain of upgrades (not just first bottleneck). Fixing one constraint often reveals 2-3 more. ## Interpretation Guide **Utilization %:** - 95-100%: Bottleneck (system constraint) - 85-95%: Constrained (near capacity) - 70-85%: Loaded (busy but buffer exists) - 50-70%: Available (spare capacity) - <50%: Underutilized (significant waste) **Bottleneck Priority:** - Utilization >95%: Immediate improvement (limiting output) - 90-95%: High priority (approaching constraint) - <90%: Not bottleneck (don't invest in capacity yet) **System Health:** - One bottleneck at 95%, others <80%: Healthy (buffer protects constraint) - Multiple stations >90%: Fragile (no buffer; any disruption impacts output) - All stations <70%: Excess capacity (reduce costs or increase demand) ## Practical Tips • **Measure capacity at each station** — Time studies, throughput data, theoretical calculations • **Visualize WIP accumulation** — Queue before bottleneck station reveals constraint visually • **Protect the bottleneck** — Preventive maintenance, backup parts, buffer inventory upstream • **Subordinate non-bottleneck** — Run them slower to match bottleneck pace (reduce WIP) • **Continuous elevation** — After improving bottleneck, identify new bottleneck and repeat • **Don't over-optimize** — Match capacity to demand; excess capacity beyond demand is waste • **Cross-train workers** — Flexibility to reallocate labor to bottleneck during breaks/absences ## Common Mistakes • **Optimizing non-bottleneck** — Improving fast station doesn't increase output; wasted effort • **Balancing line** — Making all stations equal capacity; creates fragile system with no buffer • **Ignoring downtime** — Theoretical capacity vs. effective (accounting for breaks, maintenance, efficiency) • **Investing everywhere** — Spreading capital across all stations; focus on bottleneck ROI is infinite • **Not measuring WIP** — Excessive WIP signals bottleneck or poor pacing • **Batch optimization on non-bottleneck** — Optimizing setup time on fast station is low-value vs. optimizing bottleneck ## When NOT to Use Bottleneck Analysis • **Job shop** — Highly variable custom orders; bottleneck shifts constantly (use drum-buffer-rope) • **Parallel processes** — Multiple independent lines; each has own bottleneck • **Service businesses** — Capacity isn't physical machine; human bottleneck is different analysis • **R&D/creative work** — Throughput optimization inappropriate for knowledge work (quality matters more than speed)

History

Production bottleneck analysis evolved from intuitive observation to systematic Theory of Constraints methodology as manufacturers discovered that optimizing overall system capacity required focusing exclusively on constraint stations rather than improving all workstations equally. ## Origins & Why It Emerged Early manufacturing (1900s-1950s) intuitively understood bottlenecks: "Assembly line moves at pace of slowest worker." But optimization was trial-and-error: add workers until output increases, then add to next slow station. No systematic framework. Time-and-motion studies (Frederick Taylor, 1911) measured each workstation: how long does each step take? This enabled identifying slowest step. But optimization remained localized: improve each station independently. The insight that improving non-bottleneck stations is wasteful wasn't formalized. The Goal (Eliyahu Goldratt, 1984) revolutionized production thinking with Theory of Constraints (TOC). Core ideas: (1) System capacity = bottleneck capacity (not sum of parts), (2) Optimize bottleneck only (improving others wastes resources), (3) Subordinate everything to bottleneck (keep it running always). This challenged conventional wisdom: "improve everything" is inefficient. Lean manufacturing (Toyota Production System, 1990s globalization) complemented TOC. Lean focused on waste elimination; TOC focused on bottleneck optimization. Together: Modern production management. ## How It Evolved in Practice Pre-TOC (1970s-1980s): Balanced lines were goal. If Assembly does 80/hr, "fix" Cutting to also do 80/hr (balanced). Goldratt showed: This wastes Cutting's capacity. Better: Let Cutting run at 100/hr (creates buffer before Assembly), ensure Assembly never stops. 1990s-2000s: TOC adoption grew. Manufacturing software (APS - Advanced Planning & Scheduling) incorporated TOC: schedule around bottleneck, buffer before constraint, release materials at constraint pace. Companies reported: 10-30% throughput gains without capex (just by focusing on constraint). 2010s-Present: Digital twins and simulation model entire production systems. Identify bottleneck under various scenarios (demand spikes, equipment failure). Predictive maintenance prioritizes bottleneck (prevent downtime there). IoT sensors monitor: bottleneck utilization in real-time (alert if dropping below 95%—investigate why). ## Modern Usage Today Modern manufacturers continuously identify and elevate bottlenecks. Process: (1) Measure capacity at each station, (2) Find minimum (bottleneck), (3) Improve (add equipment, optimize process, reduce downtime), (4) New bottleneck emerges, (5) Repeat. This endless cycle of constraint elevation drove manufacturing productivity gains. ## Common Misconceptions Historically • **"Balance line for efficiency"** — Balanced line is fragile; deliberately unbalanced with buffer at bottleneck is better • **"Optimize everything"** — Wasted effort; non-bottleneck improvements don't increase output • **"Utilization should be 100% everywhere"** — Non-bottleneck 100% utilization creates excess WIP (waste); should match bottleneck pace • **"More capacity always better"** — Capacity beyond demand is wasted investment