The Sizing Formula
Line sizing is a four-step formula that converts annual demand into a nameplate speed you can shop for. Skip any step and you will either overspend on capacity you never use or underbuy and stock out in peak season.
Required Nameplate IPM = (Annual Demand / Net Production Hours) / Target OEE x Growth Factor
Annual Demand is the units you must ship in a year, not what marketing hopes to sell. Use a rolling 12-month forecast tied to actual purchase orders and committed pipeline. Hope is not a planning input.
Net Production Hours is the hours the line actually runs product, after subtracting breaks, meals, maintenance, holidays, and changeovers from the gross shift schedule. This is always smaller than plant managers think.
Target OEE is the sustained Overall Equipment Effectiveness you expect to hit at maturity. Plan for 65-70% during the first 6 months, 72-78% after that. World-class is 85%. Most lines land at 60-75%.
Growth Factor covers Year 3 demand, demand spikes, and forecast error. Use 1.15 for mature products with stable demand, 1.25 for growing categories, 1.30 for new launches.
The formula gives you a nameplate speed in items per minute. That number drives every downstream decision: machine model, conveyor length, buffer sizing, operator headcount, and floor space. For a deeper treatment of OEE and nameplate versus sustained speed, see machine speed and throughput. For how this formula fits into the broader production line pillar, see the Production Line overview.
Available Production Hours Reality Check
Gross shift hours always overstate what a line actually delivers. A plant running two 8-hour shifts, 5 days a week, 50 weeks a year nominally has 4,000 hours of capacity. Net production hours, after realistic deductions, are typically 3,100-3,400.
Here is the breakdown for one shift of 8 hours.
Gross: 8.0 hours. Subtract 15 minutes for shift handover and pre-start checks. Subtract two 15-minute breaks. Subtract 30 minutes for meal. Subtract 15 minutes for end-of-shift cleanup and line clearance. Subtract 10 minutes average for minor jams and film splices spread across the shift. Net per shift: approximately 6.6 hours.
Across 2 shifts x 5 days x 50 weeks, that is 3,300 hours. Subtract another 4-6% for holidays, planned maintenance shutdowns, and unplanned downtime events. Realistic annual net: 3,100-3,150 hours.
The deduction matters because it flows straight into the sizing formula. Use 4,000 hours instead of 3,100 and you under-spec the line by 22%. The line then runs behind plan every week and nobody can explain why.
A second reality check: changeover time. If the line runs 4 SKUs and each changeover takes 25 minutes, and you average 3 changeovers per shift, that is 75 minutes per shift of zero output. Over a year, that is another 600 hours lost on a 2-shift line. Account for it explicitly in net hours or accept that the line will underperform.
For lines with high SKU variety, changeover can consume 15-25% of gross shift hours. SMED (Single-Minute Exchange of Die) discipline is the only practical way to claw that back without buying a second line.
OEE Buffer: Why 1.3x Nameplate Minimum
The gap between nameplate speed and sustained throughput is not a bug. It is physics. Every line loses time to micro-stops (a bag jams, a film splice, a weigher feed issue), to minor speed drops (the operator dials back to stabilize a seal), and to quality checks (rejecting a sample, clearing a minor defect).
The 1.3x rule says: buy a line whose nameplate is at least 1.3 times the sustained throughput you need. If the formula says you need to sustain 200 IPM, the nameplate should be at least 260 IPM. Here is why.
A line sized exactly to sustained demand has no headroom. Every micro-stop drops output below plan. Every changeover pushes the schedule behind. Every quality drift forces a speed reduction. The line runs permanently behind and the plant adds weekend shifts to catch up, which costs more in labor and overtime than the nameplate upgrade would have cost.
The 1.3x buffer decomposes into three parts. OEE loss accounts for roughly 25-35% (the inverse of 70-75% OEE). Growth and demand variability accounts for another 10-15%. Commissioning ramp accounts for 5-10% during the first 6 months as the line is dialed in.
Push the buffer above 1.3x and you start paying for capacity you will not use. A 1.5x line on a stable mature product is overspecified. Below 1.3x and the line cannot absorb a single bad week without falling behind.
For new product launches where demand is uncertain, 1.35-1.4x is justified because the forecast error is larger. For mature commodity products with tight margins and accurate forecasts, 1.25-1.3x is the sweet spot.
Worked Example: 50M Snack Bags Per Year
A snack food producer needs to package 50 million 50g pillow bags in 2026. The plant runs 2 shifts of 8 hours, 5 days a week, 50 weeks a year. The line runs 4 SKUs with an average of 3 changeovers per shift at 25 minutes each.
Step 1: Net production hours. Gross shift: 8 hours. Net per shift after breaks, meals, cleanup, and minor stops: 6.6 hours. Changeover loss: 75 minutes per shift, so net run time per shift: 5.35 hours. Across 2 shifts x 5 days x 50 weeks: 2,675 hours. Subtract 5% for holidays and planned maintenance: 2,540 net production hours.
Step 2: Required sustained throughput. 50,000,000 / 2,540 = 19,685 units per hour, or 328 IPM sustained.
Step 3: Apply OEE. New line, first year. Target 68% OEE. 328 / 0.68 = 482 IPM at nameplate with zero growth buffer.
Step 4: Apply growth factor. Category is growing 8% per year. Use 1.20 to cover Year 2 and Year 3 demand. 482 x 1.20 = 578 IPM.
Step 5: Round and shop. The producer should be shopping for a VFFS line rated at approximately 580-600 IPM nameplate. That typically means a high-speed single-lane VFFS or a dual-lane configuration running 290-300 IPM per lane.
The cost difference is significant. A 250 IPM VFFS runs $90,000-$140,000. A 600 IPM high-speed VFFS or dual-lane runs $220,000-$380,000. But the 250 IPM line cannot hit 50M bags in 2,540 hours no matter how well it is run. The producer either buys the right machine or accepts they will miss volume.
Multi-Product Lines: Weighted Average Sizing
Most lines do not run one product. They run 3-8 SKUs with different film, pouch size, weight target, and cycle characteristics. Sizing for the fastest SKU overspends. Sizing for the slowest SKU underperforms. The correct method is a weighted average.
Calculate the sustained IPM each SKU requires (using the formula above, but on a per-SKU basis). Then weight by each SKU's share of annual runtime, not by unit volume share. Runtime matters because a low-volume SKU with a slow cycle can still eat a disproportionate share of hours.
Example. A line runs 3 SKUs. SKU A: 150 IPM, 50% of runtime. SKU B: 110 IPM, 30% of runtime. SKU C: 80 IPM, 20% of runtime. Weighted average: (150 x 0.50) + (110 x 0.30) + (80 x 0.20) = 75 + 33 + 16 = 124 IPM sustained.
Size the line for 124 IPM sustained, then apply OEE and growth factors as usual.
There is one caveat. The slowest SKU (SKU C at 80 IPM) must still be able to run on the line without becoming the bottleneck. If the line is sized for 124 IPM sustained but SKU C requires a different bag former or a slower seal cycle, the line will lose time every time SKU C runs. Either accept the loss, or specify equipment that can run all SKUs at their required speed.
For high-mix lines (8+ SKUs), the weighted average tends to drift toward the slowest SKU because changeover time dominates. In that case, the right answer is often a more flexible mid-speed machine with fast SMED changeover rather than the highest-speed machine available.
Growth Headroom: Size for Year 3
Packaging equipment has an 8-12 year service life. Demand in Year 3 is almost always higher than demand in Year 1. If you size the line for current demand, you will be shopping for an upgrade in 3 years, and a second line costs 2-3x what oversizing the first line would have cost.
The rule: size for Year 3 projected demand, capped at 25% above current. Beyond 25%, the cost of idle capacity in Years 1-2 outweighs the avoided expansion cost.
Example. Current demand: 50M units. Year 3 forecast at 12% annual growth: 70M units, which is 40% above current. Cap at 25%: size for 62.5M units. The remaining 7.5M units of Year 3 demand should be met by adding a Saturday shift, extending weekday hours, or buying a second smaller line in Year 2.
Growth headroom interacts with the OEE buffer. A line with 1.3x nameplate buffer and 1.2x growth factor has 1.56x total headroom. That is appropriate for a growing category. A line with 1.3x nameplate and 1.0x growth has 1.3x total, appropriate for a mature flat category.
The mistake to avoid: inflating both factors because of optimism. A 1.4x nameplate with 1.35x growth yields 1.89x total headroom. On a 50M unit demand, that is sizing for 95M units. Unless Year 3 forecast actually supports that number, the line will run at 55% utilization for years and the per-unit cost will be 30% higher than necessary because depreciation is spread over fewer units.
Common Sizing Mistakes
Using gross shift hours. The single most common error. Gross hours overstate capacity by 20-30%. Always size on net production hours after breaks, meals, cleanup, and changeover.
Trusting vendor demo OEE. Vendors demo at 85% OEE in a controlled environment with their best operator, their product, and ideal film. Plan for 65-70% in your plant for the first 6 months. If you bake 85% into the sizing formula, the line is 20% undersized on day one.
Ignoring changeover time on multi-SKU lines. Changeover is invisible on a single-SKU line. On a 6-SKU line, it can consume 20% of gross hours. Account for it in net hours or the line underperforms.
Sizing for the fastest SKU. Sales teams love quoting the fastest SKU because it justifies a bigger machine. The weighted average method gives you the right number. The fastest SKU is the wrong anchor.
Sizing for current year only. Equipment lasts a decade. Demand grows. If you do not add 15-25% growth headroom, you will be shopping again in 3 years at a 2-3x cost premium.
Sizing for peak month. Peak month demand can be 30-40% above average. Sizing the line for peak means it runs at 70% utilization the rest of the year. The right answer is to size for average-plus-growth and meet peak demand with extended shifts or weekend production. See the FAQ on weekend production above.
Forgetting floor space. A 600 IPM line is physically larger than a 250 IPM line. The bigger line needs more conveyor length, more buffer space, more operator aisle room. If the floor cannot fit it, the right-sized machine on paper becomes an impossible install in practice.
Treating OEE as a constant. OEE varies by SKU, by shift, by operator, by product age. A line might run SKU A at 78% OEE and SKU C at 62% OEE. Use SKU-specific OEE in the weighted average or the result is misleading.
Ignoring the bottleneck station. The slowest station sets line speed. A 600 IPM filler paired with a 350 IPM checkweigher delivers 350 IPM, not 600. Balance the line before sizing the headline number. For more on this, see line balancing and bottlenecks and the deeper treatment of buffer between stations in buffer and accumulation.
Sizing is the first and most consequential decision in line design. Get the formula right, get the net hours right, get the OEE assumption honest, and the line will deliver. Get any of them wrong, and no amount of operational excellence will recover the gap.