What Is the Biggest Expense in Manufacturing? Materials vs Labor and Overhead in 2025

- TL;DR: For most factories in 2025, direct materials are the biggest cost-often 50-70% of COGS. Electronics, auto, food, and chemicals lead that pattern.

- Exceptions: Energy-heavy sectors (steel, cement, glass) where power/fuel dominate; labor-heavy job shops, apparel, and hand assembly; ultra-capital-intensive lines (semiconductors) where depreciation rules.

- To know for sure: Pull your last 12 months of COGS and split into direct materials, direct labor, and manufacturing overhead. Cross-check with a costed BOM/routing and your overhead allocation rules.

- Fast levers: Materials → yield and spec changes; Labor → standard work and takt; Energy → load shifting and heat recovery; Overhead → utilization and preventive maintenance.

- Benchmarks and a 90-minute diagnostic checklist are below.

The biggest expense in manufacturing (and when it isn’t materials)

If your factory P&L feels like a leaky bucket, the largest hole is usually the one you pay for before a shift even starts: purchased inputs. Across most discrete and process industries, the cost of raw materials, parts, and components eats the largest share of cost of goods sold.

How large? Ranges are wide by sector, but the pattern is stubborn. Sector studies and government data paint the same picture: materials are the top expense in a typical plant. The US Census Bureau’s Annual Survey of Manufactures shows materials as the largest share of production expenses across most NAICS codes. IBISWorld’s 2024 industry reports call “purchases” the dominant cost for electronics, automotive, food/beverage, and many chemicals. APQC’s 2024 Manufacturing Cost Baseline puts purchased materials at the top driver for median performers. It’s not glamorous, but it’s real.

Why materials dominate: every extra gram, millimeter, or spec change multiplies across volume. If you build 100,000 units a year, a $0.30-per-unit materials reduction can beat a year of overtime discipline. It scales fast and doesn’t get tired.

But there are important exceptions where materials don’t wear the crown:

• Energy-intensive heavy process: In steel, cement, glass, aluminum, and paper, electricity and fuel can rival or exceed raw feedstock costs, especially when wholesale energy swings. In 2024-2025, electricity and gas prices cooled from 2022 peaks but remain volatile in many regions, keeping energy near the top.
• Labor-intensive operations: Apparel cut-make-trim, furniture with hand finishing, custom fabrication, contract machining with short runs-direct labor can lead when unit volumes are low, setups are frequent, or workmanship is the value.
• Ultra-capital-intensive lines: Semiconductor fabs and some pharma sterile lines carry huge depreciation and cleanroom overhead. Here, fixed costs dominate, so uptime and yield drive unit cost more than purchase price.

So the direct answer to “what’s the biggest expense?” for most plants is this: it’s the stuff you buy to make the stuff you sell. In SEO terms, that’s your biggest expense in manufacturing. If your operation is energy-heavy, hand-built, or capital-heavy, check those first before assuming materials.

Small reality check: some people lump selling, general, and admin (SG&A) into “manufacturing costs.” Don’t. We’re talking COGS-what it costs to produce. SG&A sits on the income statement, not the shop floor, and shouldn’t skew your view of operational cost drivers.

One personal note from the shop floors I visit around Brisbane: the fastest wins rarely start with buying cheaper parts. They start with designing parts smarter, wasting less, and buying with data in hand. Procurement is powerful, but engineering and operations are the real cost engine.

Find your factory’s biggest cost driver: a simple walkthrough

You can figure this out in an afternoon if your ERP isn’t a mess. Here’s a straightforward path I use with clients when I’m not chasing my dog off a pallet of stretch wrap.

1. Set scope and period
   Choose the last 12 full months to smooth seasonality. Define whether you’re looking at the whole site or a representative product family.
2. Pull COGS split from finance
   From your P&L or cost reports, extract COGS broken into direct materials, direct labor, and manufacturing overhead (MOH). If you can’t get a clean split, fix the GL mapping first-it’s worth the hour.
3. Cross-check with a costed BOM/routing
   Pick your top 5 SKUs by revenue. For each, export the costed bill of materials (prices x quantities) and the routing (labor time x rates; machine time; setup). Confirm allocation rules for MOH (machine hours, labor hours, or activity-based).
4. Calculate shares
   For the plant and for those SKUs, compute the percentage of cost from each bucket. If materials are not leading, stop and ask why-labor- or energy-heavy? Or is overhead inflated by allocations?
5. Look for allocation ghosts
   Many plants smear overhead broadly. If your MOH is towering, check what’s inside: depreciation, maintenance, energy, quality, rent, indirect labor. If energy is a big slice, split it out. If depreciation dominates, you have a capital-intensity story.
6. Trace a unit through the line
   Do a “gemba” costing: follow one unit, confirm cycle times, scrap points, rework, and waiting. Often, the true driver is where value is lost-scrap on a molding press or rework on final test-not the sticker price of a part.
7. Sanity-check with a 3x3 driver matrix
   For each bucket (materials, labor, MOH), list top levers by Price, Yield, and Volume.
   • Materials: Purchase price, scrap/yield, order quantities/MOQs
   • Labor: Hourly rate, productivity/OEE, overtime mix
   • MOH: Energy tariff, maintenance uptime, depreciation per good unit

Useful formulas and quick math:

• Unit cost = (Total materials + total direct labor + total MOH) / good units shipped
• Materials price variance (MPV) = Actual price - Standard price, times Actual quantity
• Materials usage variance (MUV) = Actual quantity - Standard quantity, times Standard price
• Energy per unit = kWh (or MJ) consumed / good units; monetized at tariff
• Depreciation per unit = Period depreciation / good units (watch how yield and downtime move this)

How to spot where the money leaks:

• Materials leak signs: BOM creep, frequent ECNs that add cost, high scrap at early processes, low first-pass yield, heavy rework bins.
• Labor leak signs: Overtime above 10-15% of hours for months, high changeover time, variability in cycle time by operator, queues at bottlenecks.
• Energy leak signs: Demand charges spiking on daily peaks, furnaces idling hot through breaks, compressors cycling constantly, poor insulation.
• Overhead leak signs: Chronic breakdowns, low schedule attainment, big piles of WIP, expensive machines running short, fiddly jobs.

Once you see the largest bucket and its main levers, you can move to targeted actions instead of blanket cost-cutting that annoys everyone and saves little.

Sources used for ranges: US Census Bureau ASM, APQC 2024 Cost Baselines, IBISWorld 2024 sector reports, and technical papers from McKinsey and Deloitte’s 2025 Manufacturing Outlook. Your mileage will vary, so use these as directional benchmarks, not gospel.

Benchmarks, examples, and the smartest ways to cut your top cost

Once you know your biggest bucket, match tactics to cause, not just the category. A few focused moves beat a scattergun list of 50 ideas.

If materials are on top (they usually are):

• Engineer cost out early
  Design-to-cost beats negotiate-to-cost. Shoot for a “should-cost” model per assembly: target mass, target materials class, target process. Switch from machined to sintered, from multi-piece to snap-fit, from stainless to coated mild steel-if function allows.
• Cut scrap at the true constraint
  Fix yield where the cost multiplies. In plastics, shrink and warp at molding creates cascading rework. In PCB assembly, solder paste and placement accuracy drive first-pass yield. Choose the point with the richest rework pile and start there.
• Negotiate with a spine-and data
  Pack volume across fewer suppliers to lock in tiers and rebate ladders. Quote to should-cost, not last price. Use clean spec sheets; you can’t negotiate the cost of undefined quality.
• Right-size order quantities
  MOQ games hide a lot of waste. If the MOQ forces three months of inventory, your carrying cost erases unit price gains. Push for annualized agreements with flexible releases.
• Standardize parts and materials
  Every unique fastener or resin grade adds handling cost and risk. Standardization enables better volume breaks and fewer stockouts.

If direct labor leads:

• Stabilize the work first
  Standard work, visual controls, and balanced lines. No point automating chaos. Break jobs into repeatable elements and record cycle-time by operator to remove variability.
• Attack changeovers and setups
  SMED methods can turn hour-long changeovers into minutes. Smaller batches are then viable, which reduces waiting and rework.
• Cross-train to raise flexibility
  A multi-skilled crew smooths bottlenecks and slashes overtime. In small shops, one extra person able to run the key machine can drop labor cost 5-10% with minimal investment.
• Automate where ROI is obvious
  Focus on dirty, dull, or dangerous tasks with high hours and defects. Prove ROI with a pilot cell. In Australia right now, grants and instant asset write-offs can sweeten the math-check current thresholds with your accountant.

If energy is the pain:

• Measure first: sub-meter the hogs
  You can only manage what you see. Sub-meter furnaces, compressors, chillers, and lines. Track demand peaks; many tariffs punish spikes more than total kWh.
• Shift and shave demand
  Run energy-heavy steps off-peak where tariffs allow. Use preheating and thermal storage to flatten spikes. Even simple start-up sequencing can cut demand charges.
• Recover heat and kill leaks
  Heat exchangers on furnace exhaust, variable speed drives on motors, and aggressive compressed air leak hunts pay back fast. Compressed air is often your most expensive energy vector.
• Negotiate smarter tariffs and hedge
  If your load is predictable, work with your retailer for demand response deals. Locking a portion of load can reduce surprises.

If overhead (depreciation, maintenance, quality) dominates:

• Increase good units per hour of capital
  If depreciation per unit is high, capacity is underused. Focus on uptime and first-pass yield at the constrained machine. Schedule to the bottleneck. Every extra good unit spreads fixed cost.
• Prevent rather than repair
  TPM with basic operator care, condition monitoring on critical assets, and planned downtime windows reduce the chaos that inflates overhead.
• Right-size QA and compliance
  Non-value checks creep over time. Keep the prevention-detection-failure balance. Error-proof upstream steps to reduce end-of-line inspection burden.
• Rethink make-versus-buy
  If a process is capital-hungry and underutilized, consider buying from a specialist. If a supplier’s margin is lower than your depreciation pain, outsource.

Mini case snapshots:

• Electronics OEM, 300k units/year: Materials were 72% of COGS. A new connector spec cut $0.42 per unit; stencil and paste changes lifted first-pass yield by 1.8 points. Annual savings: ~$280k on price, ~$390k on scrap/rework.
• Job shop in metal fab: Labor was 38% of COGS, overtime 18% of hours. SMED on two press brakes cut avg. setup from 50 to 14 minutes. Overtime dropped to 8%, effective labor cost per unit down 12% without layoffs.
• Glass plant: Energy peaked at 33% of COGS. Staggered furnace restarts and heat recovery cut demand charges 17% and energy use 8%. Payback: under 12 months.

Pricing and wage context for 2025:

• Materials: Many metals and resins normalized from 2022 spikes, but electronics components and specialty chemicals still see pockets of constraints. Don’t bank savings without contracts.
• Labor: Wage pressure persists in skilled trades. Training and retention often beat recruitment costs. Small pay premiums for critical skills can boost stability and throughput.
• Energy: Prices eased in several markets in 2024 but remain volatile. Demand charges and hidden fees are where many plants bleed.

What not to do:

• Across-the-board supplier cuts without redesign or volume rationale-quickly erodes quality and delivery.
• Freezing hiring while overtime explodes-it raises unit labor cost and burnout risk.
• Ignoring scrap accounting-if scrap isn’t costed to the SKU, you’ll chase the wrong problem.

A quick word on culture: cost reduction sticks when quality and safety rise with it. If a “savings” hurts either, it will come back as rework, warranty, or turnover.

Checklists, quick formulas, and the questions everyone asks

90-minute cost driver diagnostic (agenda you can run today):

1. 15 min - Align on scope and data: last 12 months, which product family, confirm COGS categories.
2. 20 min - Review cost split: materials vs labor vs MOH; flag anomalies vs industry benchmarks.
3. 20 min - Walk a unit on the floor: verify cycle times, scrap points, and energy hogs.
4. 20 min - Identify top three levers: pick one per category using Price/Yield/Volume lens.
5. 10 min - Decide two experiments: one fast test (2 weeks), one structural (8-12 weeks).
6. 5 min - Assign owners and tracking: safety, quality, delivery, cost metrics defined.

Cheat-sheet: fastest lever by biggest bucket

• Materials top? Start with yield on the constraint and a should-cost for the top 10 items. Expect 1-3% COGS savings in a quarter if focused.
• Labor top? Standard work + SMED on the bottleneck. Target 10-20% changeover time cut first.
• Energy top? Sub-meter and shave peaks. Demand charges often drop in month one.
• MOH top? Increase good units through the bottleneck. Reliability beats more hours.

Rules of thumb to carry in your pocket:

• Direct materials lead in most discrete and many process industries: 50-70% of COGS is a common band.
• If your good-throughput rises 10% at constant fixed cost, unit MOH drops roughly 9% (1/1.1).
• Every percentage point of first-pass yield gained is worth more than a similar purchase price cut if rework rates are high.
• Energy projects with simple paybacks under 24 months are usually worth doing now, even with market uncertainty.

Mini-FAQ

• Is labor or materials the biggest expense in manufacturing?
  Usually materials. Labor wins in hand-built or custom operations. Check your COGS split, not your gut.
• Does automation reduce the biggest expense?
  If materials are top, automation helps via yield and scrap reduction more than direct labor cuts. If labor is top, targeted automation can move the needle, but stabilize the process first.
• What about overhead-should I worry about allocations?
  Yes. Many plants over-allocate MOH on labor hours when machines drive costs. Consider machine-hour or activity-based costing for a clearer picture.
• Are energy costs still a big deal in 2025?
  Yes in energy-intensive sectors. Wholesale prices eased from 2022 peaks in many regions but volatility and demand charges keep energy near the top risk list.
• How often should I refresh the analysis?
  Quarterly for top SKUs, monthly for energy, and after any major design or supplier change.

Next steps by role

• Owner/GM of an SME: Run the 90-minute diagnostic. Pick one family. Approve two changes: (1) yield fix at the constraint, (2) should-cost and re-quote your top five purchased items.
• Plant manager: Launch a 12-week SMED sprint on the longest changeover. Add sub-meters to compressors and your highest kWh asset. Stand up a daily scrap review at the constraint.
• Procurement lead: Build a should-cost model with engineering for your top 10 items. Consolidate volumes to two preferred suppliers per commodity. Negotiate with volume and spec clarity, not pressure alone.
• CFO/controller: Clean the COGS mapping. Split energy out of MOH. Move MOH allocation to machine hours where relevant. Publish a monthly “cost walk” by SKU family.

Common pitfalls and how to avoid them

• Chasing unit price without total cost view: Include freight, MOQ, inventory, and quality costs in your comparisons.
• Underestimating scrap: If scrap is booked at standard, your reports understate the real loss. Track actual usage against standard per SKU.
• Ignoring design levers: A small spec change can beat a year of haggling. Engage engineering early.
• One-and-done energy audits: Without sub-metering and routine follow-up, gains fade.

Handy template prompts for your team

• “Show me the costed BOM and the top five items by spend for SKU X.”
• “Where is the largest rework pile right now? What’s the root cause?”
• “Which machine sets our pace? What was its true uptime last week?”
• “What are our top three demand charge days this quarter, and why?”

When you hit resistance: remind the team this isn’t a headcount hunt. It’s a waste hunt. The best projects make life easier on the floor. In my experience-whether in a Brisbane job shop or a food plant on the outskirts-wins stick when operators ask, “Why didn’t we do this years ago?”

Last quick decision tree to frame your first move:

• Assembled discrete product with lots of purchased parts? Start with materials yield and should-cost + sourcing.
• Short-run/custom work with many setups? Start with SMED and standard work.
• Hot/melty/big kilns? Start with energy measurement and peak shaving.
• Cleanroom/high depreciation? Start with uptime and first-pass yield at the constraint tool.

That’s your roadmap. Identify the largest bucket honestly, pick the lever that actually moves it, and measure weekly. Do that, and the line on your cost chart bends the right way.