Reference

Manufacturing terms with formulas and calculators.

A practitioner's glossary of manufacturing terms: OEE, takt time, Cpk, MRR, chip load, MRP, EOQ, and more, each with a definition, formula, and calculator.

Overall Equipment Effectiveness (OEE)

  • A single percentage that measures how much of the planned production time is truly productive. It multiplies three factors: availability (uptime versus planned time), performance (actual speed versus rated speed), and quality (good parts versus total parts). A world-class OEE is around 85 percent; most plants run 45 to 65 percent.
  • OEE = Availability x Performance x Quality

Takt Time

  • The pace at which you must complete one unit to meet customer demand, calculated as available production time divided by demand. If a line has 27,000 seconds available in a shift and must build 450 units, takt time is 60 seconds per unit. It sets the drumbeat every station must match.
  • Takt = Available time / Customer demand

Cycle Time

  • The actual time it takes to produce one unit at a process, measured from start to start of consecutive units. Cycle time must be at or below takt time to keep up with demand. It excludes waiting and queue time, which are counted separately in lead time.

Lead Time

  • The total elapsed time from order release to completion, including queue, setup, run, wait, and move time. In most shops value-added run time is only 10 to 20 percent of lead time, with the rest spent waiting, which is why lead-time reduction focuses on queues rather than machine speed.

Process Capability (Cpk)

  • A measure of how well a process fits inside its specification limits, accounting for centering. A Cpk of 1.33 is a common minimum for capable processes and corresponds to roughly 63 defective parts per million; 1.67 is often required for safety-critical features.
  • Cpk = min[(USL - mean), (mean - LSL)] / (3 x sigma)

Sigma Level

  • A count of how many standard deviations fit between the process mean and the nearest specification limit, used to express defect rate. Six Sigma corresponds to 3.4 defects per million opportunities once the standard 1.5 sigma long-term shift is included.

First Pass Yield (FPY)

  • The share of units that pass every step correctly the first time, with no rework or scrap. It is the product of the yield at each step, so a five-step line at 98 percent each yields only about 90 percent overall. FPY exposes hidden rework that final yield hides.

Scrap Rate

  • The percentage of produced material or parts discarded as unusable. It is tracked by cause code rather than as a single number, because setup scrap, defect fallout, and kerf loss each respond to different fixes. Scrap cost includes material, labor, and machine time already invested.

Material Removal Rate (MRR)

  • The volume of material a cutting tool removes per unit time, in cubic inches or cubic centimeters per minute. For milling it is width of cut times depth of cut times feed rate. MRR drives spindle horsepower demand and roughing cycle time.
  • MRR (milling) = Width x Depth x Feed rate

Surface Feet per Minute (SFM)

  • The speed of the cutting edge relative to the workpiece, the anchor value for every speeds-and-feeds calculation. It comes from the tool and material pairing, not the machine. Carbide in aluminum runs 800 to 1000 SFM; the same tool in stainless drops to 150 to 300 SFM.
  • RPM = (SFM x 3.82) / diameter

Chip Load

  • The thickness of material each cutting edge removes per revolution, also called feed per tooth. It is the real design variable behind feed rate. Too light a chip load rubs and work-hardens the material; too heavy chips the edge. Typical values run 0.001 to 0.006 inch depending on tool diameter.
  • Feed rate = RPM x flutes x chip load

Kerf

  • The width of material removed by a cutting process such as sawing, laser, plasma, or waterjet. Kerf must be accounted for in nesting and cut lists because it becomes scrap on every cut. A 0.06 inch kerf across hundreds of cuts adds up to real yield loss on a sheet.

Draft Angle

  • The taper designed into the walls of a molded or cast part so it releases cleanly from the tool. A common rule is one degree of draft per inch of depth, with textured surfaces needing more. Too little draft causes drag marks, ejector-pin stress, and longer cycle times.

Clamp Tonnage

  • The force an injection molding or stamping press applies to keep the tool closed against internal pressure. For injection molding it is estimated as projected part area times cavity pressure, typically 2 to 5 tons per square inch. Undersized tonnage causes flash; oversized wastes energy and machine capacity.

Geometric Dimensioning and Tolerancing (GD&T)

  • A symbolic language on engineering drawings that defines allowable variation in form, orientation, location, and runout relative to datums. GD&T communicates function precisely, so a hole is toleranced by true position rather than by two independent coordinate dimensions.

First Article Inspection (FAI)

  • A full dimensional and material verification of the first part from a new or changed production process, documented against every drawing requirement. FAI proves the process can make a conforming part before a full run is authorized, and is a standard requirement in aerospace and automotive.

Production Part Approval Process (PPAP)

  • The automotive-standard package a supplier submits to prove it can consistently make a part to specification, including the FAI, control plan, process capability studies, and measurement system analysis. Approval is required before a supplier ships production volumes.

Mean Time Between Failures (MTBF)

  • The average operating time between failures of a repairable asset, a core reliability metric. A higher MTBF means fewer breakdowns. It is paired with MTTR to model availability and to size spare-parts inventory and maintenance staffing.
  • MTBF = Total operating time / Number of failures

Mean Time to Repair (MTTR)

  • The average time to restore a failed asset to service, including diagnosis, repair, and testing. Lower MTTR raises availability. Together MTBF and MTTR give availability as MTBF divided by the sum of MTBF and MTTR.
  • Availability = MTBF / (MTBF + MTTR)

Material Requirements Planning (MRP)

  • A method that explodes a production schedule through the bill of materials to calculate what components to buy or build and when. It nets gross requirements against on-hand inventory and scheduled receipts, then offsets by lead time so material arrives just before it is needed.

Economic Order Quantity (EOQ)

  • The order size that minimizes the combined cost of ordering and holding inventory. It rises with demand and order cost and falls as holding cost increases. EOQ is the classic trade-off between ordering too often and carrying too much stock.
  • EOQ = sqrt(2 x demand x order cost / holding cost)

Reorder Point (ROP)

  • The inventory level that triggers a replenishment order, set to cover demand during the supplier lead time plus a safety buffer. With demand of 40 units per day and a 12 day lead time, the base reorder point is 480 units before safety stock.
  • ROP = (Daily demand x Lead time) + Safety stock

Safety Stock

  • Extra inventory held to absorb variability in demand and supply lead time and to protect a target service level. It is often set as a service factor times the standard deviation of demand over lead time; a 95 percent service level uses a factor of about 1.65.

Work in Process (WIP)

  • The inventory of partially finished units between the start and end of production. By Little's Law, average WIP equals throughput times lead time, so cutting WIP shortens lead time at a fixed throughput. Excess WIP hides quality problems and ties up cash.
  • WIP = Throughput x Lead time

Throughput

  • The rate at which a system produces finished goods, set by its slowest resource, the bottleneck. Improving a non-bottleneck station does not raise throughput. It is measured in units per hour or per shift and is the numerator of most productivity metrics.

Bottleneck

  • The resource with the least capacity in a process, which sets the throughput of the whole line. An hour lost at the bottleneck is an hour lost for the entire system, so scheduling, buffering, and improvement effort concentrate there first.

Changeover Time

  • The time to switch a machine or line from making the last good part of one product to the first good part of the next. SMED techniques target reducing it below ten minutes by converting internal setup steps, done while stopped, into external steps done while running.

Single-Minute Exchange of Die (SMED)

  • A lean method for cutting changeover time to under ten minutes by separating internal setup, which requires the machine stopped, from external setup, which can be prepared in advance. Faster changeovers enable smaller batches and lower inventory.

Design for Manufacturability (DFM)

  • The practice of designing parts so they are easy and cheap to make, by simplifying geometry, relaxing non-critical tolerances, standardizing features, and choosing processes early. Most of a part's cost is locked in during design, so DFM has the highest leverage on unit cost.

Design for Assembly (DFA)

  • The practice of designing products so they are quick and error-proof to assemble, by reducing part count, adding self-locating features, and enabling one-direction assembly. Fewer parts means fewer purchase orders, less inventory, and lower labor time.

Kaizen

  • The practice of continuous, incremental improvement driven by the people who do the work. Rather than large projects, kaizen makes many small changes to eliminate waste, often in short focused events on the shop floor.

Kanban

  • A pull-based signaling system that authorizes production or replenishment only when downstream demand consumes stock. A card, bin, or electronic signal caps work in process and prevents overproduction, keeping inventory tied to actual consumption.

Poka-Yoke

  • A mistake-proofing device or design that makes an error impossible or immediately obvious, such as a fixture that only accepts a part in the correct orientation. It targets the defect at its source rather than catching it later in inspection.

Andon

  • A visual signaling system, often a light or board, that lets any operator flag a problem and, when needed, stop the line so it is fixed immediately. Andon makes abnormalities visible in real time rather than letting defects flow downstream.

Jidoka

  • The principle of building in the ability of a machine or operator to stop automatically when a defect or abnormality occurs, so bad parts are not passed on. It separates human work from machine work and prevents mass production of defects.

Value Stream

  • The full set of steps, both value-adding and wasteful, required to bring a product from raw material to the customer. Value stream mapping documents material and information flow to expose where lead time and inventory accumulate.

Yield

  • The fraction of input material or units that becomes good, sellable output. For material-heavy processes yield is dominated by nesting and scrap; for assembly it is dominated by defect and rework rates. Every point of yield is direct margin.

Uptime

  • The percentage of scheduled time that equipment is available and running, the availability factor inside OEE. Downtime is split into planned events like changeovers and unplanned events like breakdowns, each tracked separately to target the right fix.

Feed Rate

  • Feed rate is how fast the tool advances through the work, in inches per minute for milling or inches per revolution for turning. Get it wrong and you either burn tools or waste spindle time. Start from the insert maker's chip load, typically 0.001 to 0.005 inch per tooth for small end mills, and adjust from there.
  • Feed rate (ipm) = RPM x number of flutes x chip load

Depth of Cut

  • Depth of cut is how deeply the tool engages the workpiece per pass, axially or radially. Along with speed and feed it sets material removal rate and cutting force. Roughing passes in steel often run 0.100 to 0.250 inch deep; finishing passes drop to 0.005 to 0.020 inch to hold size and surface finish.

Runout

  • Runout is the wobble of a rotating tool or part, measured with an indicator as total indicator reading (TIR). Every thousandth of runout at the tool tip shows up as uneven chip load, oversize holes, and chipped edges. A good CNC toolholder setup keeps runout under 0.0005 inch; past 0.001 inch, expect tool life to fall sharply.

Backlash

  • Backlash is the lost motion in a screw or gear drive when it reverses direction. On a worn manual mill it can reach 0.005 to 0.015 inch, which is why climb milling on old iron is risky. Ball screw CNC machines hold backlash near 0.0001 inch and let the control compensate for the rest.

Surface Roughness (Ra)

  • Surface roughness Ra is the arithmetic average deviation of the surface profile from its mean line, in microinches or micrometers. A standard machined finish is 63 to 125 microinch Ra; sealing and bearing surfaces often need 16 to 32; grinding reaches 8 or better. In turning, feed rate and insert nose radius set the theoretical floor.
  • Ra = f^2 / (32 x r) (ideal turning finish, f = feed per rev, r = nose radius)

Tolerance Stack-Up

  • Tolerance stack-up is the accumulation of individual feature tolerances across an assembly. Add every tolerance in the chain worst case and you often prove the parts cannot assemble, which is why designers use statistical (RSS) stacks; with five or more contributing dimensions, RSS typically predicts 30 to 40 percent less variation than worst case. Check the stack before quoting tight print tolerances.
  • Worst case: T_total = T1 + T2 + ... ; RSS: T_total = sqrt(T1^2 + T2^2 + ...)

Datum

  • A datum is the theoretically exact plane, axis, or point that a part's features are measured from, established by physical surfaces called datum features. On a print, datums A, B, and C typically lock all six degrees of freedom. Pick datums that match how the part is fixtured and how it functions, or machining and inspection will argue forever.

True Position

  • True position is the GD&T control for how far a feature's actual location deviates from its theoretically exact location, expressed as a diametral zone. A position callout of 0.010 inch lets the hole axis sit anywhere inside a 0.010 inch diameter circle. Bonus tolerance from MMC modifiers often saves parts that a plain coordinate check would scrap.
  • True position = 2 x sqrt(deltaX^2 + deltaY^2)

Workholding

  • Workholding is everything that locates and clamps a part against cutting forces: vises, chucks, collets, vacuum tables, and dedicated fixtures. Rigidity here sets your ceiling on speed, feed, and tolerance; chatter often traces back to the clamp, not the tool. Locate on three solid points, clamp against fixed stops, and never clamp on an unqualified surface.

Fixturing

  • Fixturing is the design and use of dedicated devices that locate, support, and clamp parts repeatably for machining, welding, or assembly. Good fixtures follow the 3-2-1 principle: three locating points on the primary datum, two on the secondary, one on the tertiary. Cutting a ten minute manual setup to a thirty second load often pays for the fixture in one production run.

Spindle Speed

  • Spindle speed is the rotational speed of the tool or workpiece in RPM, set from the recommended cutting speed for the tool and material pair. Carbide in mild steel likes 300 to 500 sfm; HSS drops to around 100 sfm. Small diameter tools need high RPM to reach cutting speed, which is why micro tools max out spindles.
  • RPM = (SFM x 3.82) / tool diameter in inches

Coolant Concentration

  • Coolant concentration is the ratio of oil concentrate to water in a soluble or synthetic metalworking fluid, checked daily with a refractometer. Most sumps run 5 to 10 percent. Drift below 4 percent invites rust and bacteria; above 12 percent you get foam, sticky machines, and dermatitis complaints. Top off with premix, never straight water.

Tool Life

  • Tool life is the productive cutting time before an edge wears past its limit, measured in minutes in cut or parts per edge. Taylor's equation says life falls steeply with speed: a 20 percent speed increase can roughly halve carbide tool life. Set forced tool change intervals at about 75 percent of proven life to avoid mid-lot surprises.
  • V x T^n = C (Taylor tool life equation)

Peck Drilling

  • Peck drilling is drilling a deep hole in increments, retracting the drill each peck to break chips and let coolant reach the point. It becomes standard practice once hole depth passes about 3 to 4 times diameter. A common recipe is a first peck of one diameter with shorter pecks after, using full retract in gummy materials that pack chips.

Thread Pitch

  • Thread pitch is the axial distance between adjacent thread crests. Metric threads state pitch directly, so M10 x 1.5 advances 1.5 mm per revolution; inch threads state threads per inch, and pitch is the reciprocal, so 1/4-20 advances 0.050 inch per turn. Mixing coarse and fine pitch fasteners is a classic cause of stripped and cross-threaded assemblies.
  • Pitch = 1 / TPI (inch threads)

Tap Drill Size

  • Tap drill size is the hole diameter drilled before tapping, which sets the percentage of thread engagement. The shop rule for metric is major diameter minus pitch: M10 x 1.5 taps into an 8.5 mm hole, about 75 percent thread. Dropping to 60 or 65 percent thread loses little joint strength but can nearly halve tapping torque and tap breakage.
  • Tap drill = major diameter - pitch (approx. 75 percent thread)

Climb vs Conventional Milling

  • Climb milling feeds the work in the same direction as cutter rotation, so each tooth enters thick and exits thin; conventional milling is the reverse. On rigid CNC machines climb gives better finish and up to 50 percent longer tool life, but on machines with backlash it can yank the table into the cutter. Save conventional for worn manual iron and scaly surfaces.

Stepover

  • Stepover is the sideways distance the cutter shifts between adjacent milling passes, usually programmed as a percentage of tool diameter. Roughing typically runs 40 to 70 percent; ball nose finishing drops to 5 to 10 percent to control scallop height. Halving stepover roughly quarters the scallop but doubles cycle time, so set it from the finish spec, not habit.

Shot Size

  • Shot size is the volume or weight of melt the injection unit meters and delivers each cycle, covering the parts, runners, and cushion. The sweet spot uses 20 to 80 percent of barrel capacity; below that, resin sits too long and degrades, and above it you risk losing the cushion. Size the press so every job lands in this window.

Cavity Pressure

  • Cavity pressure is the pressure the melt actually exerts inside the mold cavity, typically 3,000 to 10,000 psi at the part, well below nozzle injection pressure because of losses through the runner and gate. It drives clamp tonnage sizing, roughly 2 to 5 tons per square inch of projected area, and is the best single signal for shot-to-shot consistency.

Gate

  • The gate is the restricted opening where melt enters the mold cavity from the runner. Its size and location control fill pattern, pressure loss, weld line placement, and cosmetics. Typical edge gates run 50 to 75 percent of wall thickness, and the gate freezes first, sealing the cavity and ending packing. Fix a troubled gate before blaming the process.

Runner System

  • The runner system is the network of channels carrying melt from the sprue to the gates in a multi-cavity mold. Cold runners are scrapped or reground every shot and can equal 20 to 50 percent of shot weight on small parts; hot runners eliminate that waste but add roughly 10,000 dollars or more per drop. Unbalanced runners fill cavities unevenly.

Sink Mark

  • A sink mark is a shallow depression on a molded part surface opposite a thick section, rib, or boss, formed when the inner core keeps shrinking after the skin freezes. The classic prevention rule is keeping rib thickness at 50 to 60 percent of the nominal wall. More packing pressure and longer hold help, but geometry fixes beat process band-aids.

Weld Line

  • A weld line, or knit line, forms where two melt fronts meet and fuse, downstream of holes, inserts, or between gates. The junction may retain only 20 to 80 percent of base material strength depending on resin and temperatures, and often shows as a visible seam. Hotter melt, faster fill, and a vent at the meeting point help; moving the gate relocates it.

Flash

  • Flash is thin excess plastic squeezed out at the parting line, vents, or ejector pin clearances during injection. It signals clamp force too low for the projected area, worn shutoff surfaces, or overpacking. Vent depths around 0.0005 to 0.002 inch, resin dependent, let air escape without passing melt. Chronic flash calls for mold repair, not more trimming labor.

Short Shot

  • A short shot is a molded part that never completely filled, with geometry missing at the last point of fill or in thin ribs. Usual suspects, in rough order: insufficient shot size or cushion, trapped air from poor venting, low melt or mold temperature, and undersized gates. Intentional short shots at 90 to 95 percent fill are a standard mold trial technique.

Cooling Time

  • Cooling time is the portion of the molding cycle after packing when the part solidifies enough to eject without distortion, commonly 50 to 80 percent of the total cycle. It scales with the square of wall thickness, so doubling a wall roughly quadruples cooling. A quick estimate for many resins is 1 to 2 times the wall thickness in millimeters, squared, in seconds.

Regrind

  • Regrind is scrap plastic, sprues, runners, and trim that gets ground up and blended back into virgin resin. It cuts material cost, but every heat history degrades the polymer, so most shops cap regrind at 10 to 25 percent of the blend and keep it away from parts with tight cosmetic or structural specs.

Mold Shrinkage

  • Mold shrinkage is how much a plastic part contracts as it cools from melt to room temperature, so the toolmaker cuts cavities oversize to compensate. Amorphous resins like ABS shrink about 0.4 to 0.8 percent, while semicrystalline resins like polypropylene run 1.5 to 2 percent. Guess wrong and you scrap a hardened steel cavity.
  • Shrinkage (%) = ((Mold Dimension - Part Dimension) / Mold Dimension) x 100

Parting Line

  • The parting line is where the two mold halves meet, and it always leaves a witness mark on the part. Smart mold designers hide it on an edge or non-cosmetic surface. Mismatch between halves should stay under about 0.005 inch; anything more means flash, sink at the seam, or a rejected cosmetic part.

Ejector Pin

  • An ejector pin is a hardened steel pin that pushes the cooled part off the core when the mold opens. Pins leave small round witness marks, so place them on ribs, bosses, or hidden faces. A common print callout holds pin marks flush within 0.005 inch; deeper marks mean ejecting too hot or too fast.

Hot Runner

  • A hot runner is a heated manifold that keeps resin molten from the machine nozzle to each cavity gate, eliminating cold runners and their scrap. Figure roughly 5 to 30 percent material savings plus faster cycles, against tooling that costs several thousand dollars per drop. It usually pays off above about 100,000 parts per year.

Family Mold

  • A family mold puts cavities for two or more different parts, usually components of one assembly, in a single tool. It cuts tooling cost and guarantees matched part quantities, but filling imbalance is the classic headache: keep shot weights of the paired parts within roughly 2 to 1 or expect short shots and flash in the same cycle.

Heat-Affected Zone (HAZ)

  • The heat-affected zone is the band of base metal next to a weld that never melted but got hot enough for its microstructure to change. It is often the weakest link, prone to hardening and cracking in steels. Typical HAZ width runs 1 to 6 mm; lower heat input and proper preheat keep it narrow and tough.

Weld Penetration

  • Weld penetration is how deep the weld metal fuses into the base metal below the surface. Too shallow and the joint fails under load; too deep and you burn through thin stock. Amperage is the main driver, and a rule of thumb for MIG on steel is about 1 amp per 0.001 inch of material thickness.

Deposition Rate

  • Deposition rate is the pounds of filler metal a process lays down per arc hour, and it drives welding labor cost more than anything else. Stick runs 2 to 5 lb/hr, solid wire MIG 5 to 8, flux-cored 8 to 25, and submerged arc higher still. Match the process to the joint before chasing travel speed.
  • Deposition Rate = Weight of Weld Metal Deposited / Arc Time

Duty Cycle (Welding)

  • Duty cycle is the percentage of a 10 minute period a welding power source can run at rated output without overheating. A 300 amp machine at 60 percent duty cycle welds 6 minutes, then rests 4. Undersize the machine and thermal shutdowns kill production; most industrial sources are rated 60 to 100 percent at working amperage.
  • Duty Cycle (%) = (Arc Time / 10 minutes) x 100

Shielding Gas

  • Shielding gas blankets the molten weld pool so oxygen and nitrogen in the air cannot cause porosity and brittle welds. For MIG on carbon steel the workhorse is 75 percent argon, 25 percent CO2 at 20 to 30 cfh; TIG uses pure argon. Wind above about 5 mph strips the shield, so go flux-cored outdoors.

Weld Distortion

  • Weld distortion is the warping and shrinkage a part suffers as weld metal cools and contracts. Expect roughly 1/16 to 1/8 inch of transverse shrinkage across a butt joint in steel plate. Fight it by welding the smallest sufficient size, balancing welds about the neutral axis, sequencing from the center out, and using strongbacks or presetting.

Fillet Weld

  • A fillet weld is the triangular weld joining two surfaces at roughly 90 degrees, as in tee and lap joints, and it is the most common weld in fabrication. Size is called out by leg length; strength is figured on the throat. Overwelding hurts: bumping a 1/4 inch fillet to 5/16 adds about 56 percent more weld metal.
  • Throat = 0.707 x Leg Size

Groove Weld

  • A groove weld fills a prepared opening between two members, typically a butt joint beveled so the arc can reach the root. Complete joint penetration grooves develop full base metal strength; partial penetration is sized by depth. Typical single V preparations use a 60 to 75 degree included angle with a 1/16 to 1/8 inch root opening.

Weld Symbol

  • A weld symbol is the standardized notation on a drawing, per AWS A2.4, that tells the welder what joint, size, length, and finish is required. The core rule: information below the reference line applies to the arrow side of the joint, above the line to the other side. A flag at the bend means weld it in the field.

Spatter

  • Spatter is the molten droplets that fly out of the arc and stick to the part and fixtures. It is pure waste: grinding time, anti-spatter spray, and nozzle cleaning. Bad parameters in globular transfer can throw away 5 to 10 percent of your wire. Tighten voltage, shorten stickout, or switch to pulsed MIG to knock it down.

Gauge R&R

  • Gauge R&R is a study that splits measurement variation into repeatability (same operator, same gage) and reproducibility (between operators), telling you whether the gage or the process is producing the spread you see. The AIAG yardstick: under 10 percent of total variation is acceptable, 10 to 30 percent marginal, over 30 percent means fix the measurement system first.
  • %GRR = (GRR Std Dev / Total Variation Std Dev) x 100

Measurement System Analysis (MSA)

  • Measurement system analysis is the family of studies covering bias, linearity, stability, resolution, and gauge R&R that proves your gages can actually see the variation you are trying to control. Do it before any SPC or capability work. Rule of thumb: gage resolution should be at most one tenth of the tolerance you are measuring against.

Control Chart

  • A control chart plots a process statistic over time against limits set at plus and minus three standard deviations from the mean. Points inside the limits are common cause noise; a point outside, or patterns like 8 in a row on one side, signal a special cause worth chasing. Control limits come from the data, never from the spec.
  • UCL/LCL = Mean +/- 3 x Std Dev

SPC

  • Statistical process control uses control charts and capability indices to run a process on data instead of tribal knowledge, catching drift before it makes scrap. The core discipline is reacting to special causes and leaving a stable process alone. Most customers want Cpk of at least 1.33, meaning the spec limit sits four standard deviations from the mean.

DPMO

  • DPMO counts defects per million opportunities, normalizing quality across products of different complexity: a wire harness with 200 solder joints gets judged fairly against a simple bracket. It is the currency of Six Sigma, where 3.4 DPMO corresponds to six sigma performance. Count real opportunities honestly or the metric will flatter you.
  • DPMO = (Defects / (Units x Opportunities per Unit)) x 1,000,000

AQL

  • AQL, the acceptable quality limit, is the worst process average a sampling plan will routinely pass, and it anchors standards like ANSI/ASQ Z1.4. Typical contracts run 0.65 to 1.0 percent for major defects and 2.5 percent for minors. Remember it describes the sampling plan, not a promise that every accepted lot is that clean.

Attribute vs Variable Data

  • Attribute data is pass or fail, a count of defects from a go/no-go check; variable data is an actual measurement like 25.03 mm. Variable data carries far more information: capability studies need around 30 measured parts, while attribute plans often demand samples of hundreds to detect the same shift. Measure when you can, count when you must.

Calibration

  • Calibration compares a gage against a reference standard traceable to a national institute like NIST, then corrects or documents the error. The classic rule is a 4 to 1 test accuracy ratio: the standard should be at least four times more accurate than the gage under test. Typical intervals run 6 to 12 months, adjusted by usage and drift history.

Traceability

  • Traceability is the documented chain linking a finished part back to its material heats, lot numbers, process parameters, and gage calibrations. It is what turns a field failure into a bounded recall of one lot instead of everything you shipped that year. Aerospace and medical work demand it; smart shops keep it anywhere the paperwork cost is bearable.

Nonconformance

  • A nonconformance is any part, material, or process output that fails to meet a specified requirement. The discipline that matters is segregation and disposition: tag it, move it to quarantine, then formally decide use-as-is, rework, repair, regrade, or scrap. Untagged nonconforming stock migrating back into good inventory is how bad parts reach customers.

CAPA

  • CAPA, corrective and preventive action, is the formal system for eliminating the cause of a nonconformance (corrective) and stopping potential problems before they occur (preventive). It is required by ISO 9001 and FDA regulations. A working rule: contain within 24 hours, find root cause within 2 weeks, and verify effectiveness before closing, typically inside 30 to 90 days.

Root Cause Analysis

  • Root cause analysis is the structured hunt for the underlying reason a failure happened, not the symptom you first see. Standard tools are 5 Whys, fishbone diagrams, and fault trees. The acid test: if your root cause ends at operator error, dig further. Asking why about five times usually lands on a process or system weakness you can fix.

8D

  • 8D is the eight discipline problem solving format, born at Ford and now standard for automotive customer complaints. The disciplines run from team formation through containment, root cause, corrective action, prevention, and team recognition. The clock that matters most is D3: customers typically expect interim containment within 24 hours and the full report in 10 to 15 working days.

Fishbone Diagram

  • A cause and effect chart, also called an Ishikawa diagram, that maps possible root causes of a problem along branches feeding one defect statement. Most shops organize branches by the 6Ms: man, machine, method, material, measurement, and environment. Build it with the operators who run the process, then verify the top two or three suspected causes with data before changing anything.

5 Whys

  • A root cause technique where you ask why a failure happened, then ask why again on each answer, typically five times, until you reach a process cause you can fix. It stops teams from treating symptoms, like replacing a blown fuse without asking why it blew. Works best on single, simple failures; use a fishbone diagram when several causes interact.

Heijunka

  • Production leveling, a lean scheduling method that smooths both volume and product mix over a fixed interval instead of building whatever orders arrive that day. Rather than running Monday flat out and Friday half idle, you release work in small repeating patterns. Leveling cuts the demand spikes that inflate WIP, overtime, and changeover chaos. Most shops level over a one to four week window.

Gemba

  • Japanese for the real place, meaning the shop floor where value is actually created. A gemba walk means leaders go watch the process firsthand instead of managing from reports and conference rooms. The discipline matters: observe the work, ask questions, respect operators, and do not assign blame. Many plants schedule daily walks of 15 to 30 minutes per area.

Muda

  • Japanese for waste, meaning any activity that consumes resources without adding value the customer pays for. Lean identifies seven classic wastes: transport, inventory, motion, waiting, overproduction, overprocessing, and defects, often remembered as TIMWOOD. Overproduction is considered the worst because it hides the others. In most plants, value-added time is under 5 percent of total lead time, so the hunting ground is huge.

5S

  • A workplace organization method with five steps: sort, set in order, shine, standardize, and sustain. The goal is a shop floor where anything missing or abnormal is obvious in seconds, which cuts search time, tool loss, and safety incidents. Sustain is where most programs die, so audit weekly with a simple scored checklist. A good rule: any tool findable within 30 seconds.

Standard Work

  • The documented best-known method for a job: the takt time, the exact work sequence, and the standard in-process inventory needed to run it. It is the baseline that makes abnormality visible; without it you cannot tell improvement from luck. Written by supervisors and operators together, not engineering alone, and revised whenever a better method proves out. Time elements over 10 or more observed cycles.

One-Piece Flow

  • Moving parts through operations one at a time instead of in batches, so each piece goes straight from one process step to the next. Lead time drops dramatically: a 100-piece batch through five 1-minute operations takes over 8 hours in batch mode but about 104 minutes in flow. It also exposes defects after one part instead of one full bin.

Pull System

  • A production control method where downstream consumption triggers upstream replenishment, usually through kanban cards or empty bins, instead of a schedule pushing work in based on forecast. Nothing gets built until something is consumed, which caps WIP automatically. Kanban quantities are sized from demand during replenishment lead time plus a safety factor, commonly 10 to 20 percent.

Cellular Manufacturing

  • Arranging dissimilar machines into compact cells, often U-shaped, so a family of similar parts gets complete processing in one area. One operator can tend several machines, walking distance drops, and WIP between operations shrinks to a piece or two. Cells work best on part families sharing 70 to 80 percent of process steps; group parts using routing analysis or group technology coding.

Line Balancing

  • Distributing work elements across stations so every station's content comes as close to takt time as possible, with no station over it. An unbalanced line runs at the speed of its slowest station while everyone else waits. Balance efficiency above 85 percent is a solid target; below 70 percent means rebalancing, combining stations, or redesigning the work.
  • Balance Efficiency (%) = Total Task Time / (Number of Stations x Cycle Time) x 100

Overall Labor Effectiveness

  • The labor counterpart to OEE, measuring how much of paid workforce time produces good product. It multiplies availability (time on task versus scheduled), performance (output versus standard), and quality (good units versus total) into one percentage. It shows whether losses come from absenteeism, slow work, or rework. Most plants find OLE lands well below machine OEE, often in the 60s.
  • OLE = Availability x Performance x Quality

Preventive Maintenance

  • Maintenance performed on a fixed schedule, by calendar time or usage such as run hours or cycles, to keep equipment from failing in service. Lubrication, filter changes, belt inspections, and calibrations are typical tasks. An unplanned breakdown typically costs 3 to 10 times more than the same repair done on a planned basis, which is the whole economic case.

Predictive Maintenance

  • Maintenance triggered by measured equipment condition, using vibration, temperature, oil analysis, or motor current signatures, rather than by calendar. You repair when the data says a failure is developing, often weeks before it happens. Studies commonly credit predictive programs with cutting maintenance costs 25 to 30 percent and breakdowns around 70 percent versus reactive maintenance. It requires baseline data and trained analysts.

Condition Monitoring

  • The continuous or periodic measurement of equipment health parameters, with vibration, temperature, oil condition, and motor current being the usual four, to catch developing faults. The key concept is the P-F interval: the time between when a fault becomes detectable and when it becomes functional failure. For rolling element bearings that window is often weeks to months, which is your planning runway.

Total Productive Maintenance (TPM)

  • A plant-wide maintenance philosophy that puts routine equipment care in operators' hands through autonomous maintenance, meaning daily cleaning, inspection, and lubrication, while maintenance crews handle planned and predictive work. Built on eight pillars and measured chiefly by OEE, where world-class is around 85 percent. Expect two to three years to embed; the payoff is breakdowns cut by half or better.

Planned Downtime

  • Scheduled non-production time for changeovers, preventive maintenance, cleaning, trials, and breaks. How you classify it changes your metrics: most OEE definitions exclude planned downtime from available time, so a plant can post 85 percent OEE while running only 60 percent of the clock. Track it separately and attack changeovers with SMED, targeting setups under 10 minutes.

Failure Mode

  • The specific manner in which a component or process fails to perform its function: a bearing seizes, a weld cracks, a seal leaks, a sensor drifts. One part can have several failure modes, each with different causes, warning signs, and consequences. Naming the mode precisely, seized versus worn versus contaminated, is what makes maintenance planning and FMEA work actionable.

FMEA

  • Failure Mode and Effects Analysis, a structured review that lists each way a design or process can fail, then scores severity, occurrence, and detection on 1 to 10 scales. Traditionally the three multiply into a Risk Priority Number, with scores above roughly 100 drawing corrective action; the newer AIAG-VDA method uses Action Priority tables instead. Do it before launch, not after the recall.
  • RPN = Severity x Occurrence x Detection

Spare Parts Criticality

  • A ranking of stocked spares by the consequence of not having one when equipment fails, usually combining downtime cost, lead time, and failure likelihood. A 50 dollar bearing with a 12 week lead time on a bottleneck machine is critical; a commodity fastener is not. Classify parts into A, B, C tiers and stock critical long-lead items regardless of usage rate.

Vibration Analysis

  • Measuring machine vibration, typically velocity in mm/s, and analyzing its frequency spectrum to identify specific faults: imbalance, misalignment, looseness, and bearing defects each show distinct signatures. ISO 20816 gives severity zones by machine class; sustained readings above roughly 4.5 mm/s on standard machines warrant action. Done monthly on critical rotating equipment, it catches bearing failures weeks to months early.

Thermography

  • Infrared imaging that maps surface temperatures to find problems invisible to the eye: loose electrical connections, overloaded circuits, failing bearings, refractory loss, and stuck steam traps. Electrical surveys are the classic use; a connection running 10 C or more above similar components under similar load warrants scheduled repair, and 40 C over means fix it now. Survey electrical rooms at least annually.

Bill of Materials (BOM)

  • The structured list of every raw material, component, and subassembly needed to build one unit of a product, with quantity per parent. Multilevel BOMs nest subassemblies and drive MRP explosions, standard costing, and backflushing, so errors multiply everywhere downstream. Audit accuracy regularly; anything below 98 percent will generate shortages and inventory errors your planners cannot explain.

Routing

  • The sequence of operations required to make a part, listing each work center, setup time, run time per piece, and any tooling or inspection steps. Routings feed scheduling, capacity planning, and standard costing, so stale run times quietly corrupt all three. True them up with time studies or machine data at least annually; a 20 percent run time error means a 20 percent load error.

Backflush

  • An inventory transaction method that automatically deducts component quantities when a finished item is reported complete, using the BOM instead of manual issues at each step. It slashes transaction labor but hides problems: if BOMs are wrong or scrap goes unreported, inventory records drift fast. Use it only with BOM accuracy above 98 percent and disciplined scrap reporting.

Available to Promise

  • The uncommitted portion of inventory and planned production that order entry can promise to new customer orders without disturbing existing commitments. It is calculated per period from on-hand stock plus scheduled receipts minus booked orders. Without it, sales promises the same units twice and the shop eats the expedite. Its value depends on inventory records and the master schedule both being trustworthy.
  • ATP = On-hand Inventory + Scheduled Receipts - Committed Customer Orders

Capacity Requirements Planning

  • The MRP follow-on step that converts planned and released orders into hours of load at each work center, using routings and lead time offsets, then compares that load against demonstrated capacity. It answers whether the material plan is actually buildable. Work centers loaded past about 90 percent of demonstrated capacity will slip; reschedule, offload, or add shifts before the week blows up.

Finite Scheduling

  • Scheduling that respects real capacity limits, loading each work center only up to available hours and sequencing jobs so none overlap on the same machine, unlike infinite loading which piles work into already overloaded periods. The output is an executable sequence with realistic dates. It only pays off if setup times, run rates, and machine availability data are accurate, so clean routings come first.

Cycle Counting

  • Counting a small subset of inventory items every day and reconciling errors, instead of shutting down for one annual physical inventory. Items are grouped ABC by value: count A items monthly or quarterly, C items once or twice a year. The real goal is finding and killing error causes, not just fixing numbers. Sustained record accuracy of 95 to 98 percent is the working target.

ABC Analysis

  • Inventory classification that ranks items by annual dollar usage. A items, roughly the top 20 percent of SKUs, typically drive 80 percent of spend and get tight cycle counting and safety stock control. B items get moderate attention, C items get bulk ordering. Without it, planners waste effort counting nuts and bolts while high-value stock runs dry.

Vendor Managed Inventory

  • Arrangement where the supplier monitors your on-hand quantities and replenishes to agreed minimum and maximum levels. Common for fasteners, resin, and weld consumables. Done right it cuts stockouts and slashes purchase order volume, often by 50 percent or more. It only works with clean usage data and bin quantities the supplier can actually trust.

Landed Cost

  • The true total cost of a purchased part delivered to your dock: unit price plus freight, duties, tariffs, insurance, brokerage, and handling. Overseas quotes that look 30 percent cheaper often shrink to 5 or 10 percent after landing, and can invert once expedited freight enters the picture. Sourcing decisions made on piece price alone lose money quietly.
  • Landed Cost = Unit Price + Freight + Duties and Tariffs + Insurance + Handling

Incoterms

  • Standardized trade terms published by the International Chamber of Commerce that define where risk, freight cost, and customs responsibility transfer between buyer and seller. The 2020 edition has 11 terms, ranging from EXW, where the buyer takes everything at the seller's door, to DDP, where the seller delivers duty paid. Quoting imports without naming an Incoterm invites a freight dispute.

Consignment Inventory

  • Stock physically held at your plant but still owned by the supplier; you pay only when material is pulled for production. Good for expensive, lumpy-demand items like castings or specialty resin because it moves carrying cost off your books. The catch is discipline: consumption must be reported accurately, usually daily or weekly, or reconciliation fights start fast.

Anodizing

  • Electrochemical process that grows a controlled aluminum oxide layer from the part surface itself rather than depositing a coating. Type II decorative coatings run 5 to 25 microns; Type III hardcoat runs 25 to 100 microns with hardness rivaling hardened steel. Roughly half the film thickness grows outward, so account for it on tight tolerance features.

Electroplating

  • Depositing a metal layer such as nickel, chrome, zinc, or copper onto a conductive part using DC current in an electrolyte bath. Deposit distribution follows current density, so edges and corners plate 2 to 3 times heavier than recesses. Typical functional deposits run 5 to 25 microns. Surface cleanliness before the tank decides adhesion more than anything else.

Passivation

  • Chemical treatment, usually citric or nitric acid per ASTM A967, that strips free iron from stainless steel surfaces so the protective chromium oxide layer reforms uniformly. It adds no dimension and no visible change, which tempts people to skip it; machined stainless will rust from embedded tool steel without it. Verify with a copper sulfate or 24 hour salt spray test.

Media Blasting

  • Propelling abrasive media such as glass bead, aluminum oxide, steel grit, or plastic at a surface with compressed air or wheels to clean, deburr, or texture it. Media choice sets the result: glass bead peens and brightens, aluminum oxide cuts and etches. For coating prep, hit the anchor profile the coating spec demands, typically 1.5 to 4 mils.

Deburring

  • Removing the sharp edges and raised material left by machining, stamping, or sawing. Methods range from hand tools to tumbling, brushing, and thermal or electrochemical processes. It gets quoted as an afterthought yet can eat 10 percent or more of part cost on complex machined work. Drawings should state the required edge break, commonly 0.1 to 0.5 mm.

Vibratory Finishing

  • Mass finishing process where parts and ceramic or plastic media vibrate together in a tub or bowl, deburring and smoothing hundreds of parts at once. Cycle times run 30 minutes to several hours depending on media aggressiveness and target finish. Media-to-part ratio matters; keep it at least 3:1 by volume so parts do not ding each other.

Conversion Coating

  • Thin chemical film, such as zinc phosphate, iron phosphate, or zirconium, formed by reacting the base metal surface rather than depositing on top of it. It is the standard pretreatment under powder coat and paint, improving adhesion and slowing corrosion creep dramatically. Films are thin, typically 0.1 to 3 microns, and add no meaningful dimension to the part.

Cure Schedule

  • The time at temperature a coating chemistry needs to fully crosslink, stated as part metal temperature, not oven air temperature. A typical powder schedule is 10 minutes at 200 C part metal temperature; heavy castings may need 30 plus minutes of oven time just to get there. Undercure shows up later as poor adhesion and solvent resistance, and the eye cannot catch it.

Dry Film Thickness

  • The thickness of a coating after it has cured and all solvents or carriers are gone, measured in mils or microns with a magnetic or eddy current gauge. Most powder coat specs call for 2 to 4 mils, about 50 to 100 microns. DFT drives both performance and cost: every extra mil is wasted powder, and thin spots fail salt spray first.

Adhesion Testing

  • Verifying a coating is actually bonded to the substrate, most commonly by cross hatch tape test per ASTM D3359 or pull-off test per ASTM D4541. Cross hatch grades run 0B to 5B, with 4B or 5B usually required. Adhesion failures nearly always trace back to pretreatment or cure, so run test panels with every process change.

Riser

  • A reservoir of extra molten metal attached to a casting that feeds shrinkage as the part solidifies, so the shrink cavity ends up in the riser instead of the part. The riser must freeze after the section it feeds; a common sizing rule makes riser modulus 1.2 times the casting section's. Risers get remelted but still cost yield, often 30 to 50 percent on steel castings.

Sprue

  • The vertical channel that carries molten metal or plastic from the pouring basin or machine nozzle into the runner system. In casting, a tapered sprue keeps the falling stream from aspirating air, a top source of oxide and gas defects. In injection molding the sprue is often the thickest flow section; cold sprues eject with the shot and become regrind.

Core (Casting)

  • A sand or ceramic insert placed in the mold to form internal passages and undercuts the pattern cannot produce, like water jackets and oil galleries. Cores are made separately, usually by shell or cold box process, and must be vented and supported; core shift is a leading cause of scrapped wall thickness. Every added core adds cost, so designers should minimize the count.

Pattern

  • The replica of the part used to form the mold cavity in sand casting, built oversize to compensate for solidification shrink, typically 1 to 2 percent depending on alloy, plus machining stock and 1 to 3 degrees of draft. Pattern quality caps casting quality; a worn or poorly gated pattern will scrap parts for years before anyone questions it.

Porosity

  • Voids inside a casting, weld, or molded part caused by trapped gas or solidification shrink. Gas porosity is round and scattered; shrink porosity is jagged and clustered at hot spots. It kills pressure tightness and fatigue life, and it hides until machining or leak test exposes it. X-ray inspection per ASTM E505 grades severity, with Level 1 the tightest.

Shrinkage Defect

  • A cavity or sink formed when a section runs out of feed metal during solidification; most alloys contract 3 to 7 percent going from liquid to solid. It concentrates at heavy sections and hot spots that freeze last. The fix is directional solidification using risers, chills, and section taper so freezing moves toward the feed source. Solidification simulation catches most of these before the first pour.

Grain Structure

  • The size, shape, and orientation of the metal crystals formed during solidification and reworked by hot working and heat treatment. Finer grain raises yield strength and toughness per the Hall-Petch relationship. Grain size is rated by ASTM E112 number, where higher means finer; 5 to 8 is typical for wrought steel. Furnace overheating coarsens grain and drags properties down with it.

Quenching

  • Rapid cooling from austenitizing temperature in water, oil, polymer, or gas to trap carbon and form hard martensite in steel. Cooling rate must beat the alloy's critical rate; water quenches fastest but distorts and cracks most, so oil is the shop default for medium alloy steels. Always temper afterward, because as-quenched martensite at 60 plus HRC is too brittle to use.

Case Hardening

  • Producing a hard, wear resistant surface over a tough core, either by diffusing carbon or nitrogen into the surface (carburizing, nitriding) or by heating only the surface (induction, flame). Case depths typically run 0.2 to 2 mm with surface hardness of 58 to 64 HRC. It is the standard recipe for gears and shafts, where the surface sees contact stress and the core sees bending.

Stress Relieving

  • Subcritical heat treatment, typically 550 to 650 C for steel held about one hour per inch of thickness, that lets residual stresses from welding, machining, or forming relax without much change to the microstructure. Skip it and parts move during later machining or in service. It is cheap insurance on weldments and heavily roughed parts headed for tight tolerance finishing.

Nesting

  • Arranging part profiles on a sheet, coil, or plate to maximize material utilization before cutting or stamping. Good nesting software routinely pushes utilization from the 60s to above 80 percent, and on sheet metal work material is often 60 to 70 percent of part cost, so a few points of nest efficiency beat most cycle time improvement projects.

Buy-to-Fly Ratio

  • The weight of raw material purchased divided by the weight of the finished part, an aerospace machining metric. Machined titanium structures commonly run 10:1 or worse, meaning 90 percent of expensive stock leaves as chips. Additive manufacturing and near-net forgings attack this directly, often getting below 2:1, which is the core economic case for AM in aerospace.
  • Buy-to-Fly Ratio = Raw Material Weight / Finished Part Weight

Near-Net Shape

  • Producing a blank so close to final geometry that only light finish machining remains, using precision casting, forging, powder metallurgy, or additive processes. The payoff is less material bought and fewer machining hours; moving from bar stock to a near-net forging can cut machining time 50 percent or more. Weigh that against higher tooling cost and minimum order quantities.

Springback

  • The elastic recovery of sheet metal after the bending load releases, opening the bend angle beyond the die angle. Higher strength materials spring back more; mild steel might recover 1 to 2 degrees while advanced high strength steel can exceed 10. Compensate by overbending, coining the bend line, or tuning die geometry, and expect it to shift from coil to coil.
  • Springback Factor K = Final Bend Angle / Initial Bend Angle

Bend Allowance

  • Bend allowance is the arc length of the neutral axis through a bend, added to the flat flange lengths to size the blank. Get it wrong and holes land off location after forming. For a 90 degree air bend in mild steel with inside radius equal to thickness, it comes out near 2.2 times the material thickness.
  • BA = (pi / 180) x bend angle x (inside radius + K x thickness)

K-Factor

  • The K-factor locates the neutral axis as a fraction of material thickness measured from the inside bend surface, and it feeds the bend allowance math that sets flat blank size. Typical values run 0.30 to 0.50, with 0.44 the common default for air-bent mild steel. Tighter radii and softer materials pull the value lower, so verify with a test bend.
  • K = t / T (neutral axis depth / material thickness)

Grain Direction

  • Grain direction is the orientation of the elongated crystal structure created when sheet is rolled at the mill. It matters because bends running parallel to the grain crack far more easily, especially in hard tempers and aluminum. Standard practice is to bend across the grain, and a bend along the grain needs roughly twice the inside radius of one across it.

Burr

  • A burr is the raised sliver of metal left on the die-side edge of a sheared or punched part. Burr height is the first visible sign of punch wear and excess die clearance; most stamping specs cap it at 10 percent of stock thickness. Beyond that, regrind the punch, because tall burrs cut hands, block assembly, and start cracks in later bends.

Piercing

  • Piercing is punching a hole in sheet stock where the slug is scrap and the hole is the product, the reverse of blanking. Keep pierced hole diameter at or above stock thickness or the punch will snap, and set die clearance around 5 to 10 percent of thickness per side. Tonnage comes straight from cut perimeter, thickness, and shear strength.
  • Piercing force = cut perimeter x material thickness x shear strength

Progressive Die

  • A progressive die strings multiple operations such as piercing, forming, and cutoff at stations spaced one pitch apart, with coil stock advancing every press stroke. A finished part drops each hit, so run rates of 30 to 800 strokes per minute are routine. Tooling can cost six figures, so progressives usually pay off above roughly 100,000 parts per year.

Manufacturing Unit Cost

  • The fully loaded cost to produce one unit: material, direct labor, machine time, scrap allowance, and allocated overhead. Material typically runs 40 to 70 percent of unit cost in discrete manufacturing. Tracking it per part number monthly, against the quoted number, is the fastest way to find margin leaks.

Direct Labor Cost

  • The wages, taxes, and benefits of the people who physically make the product, expressed per hour or per unit. Loaded rates typically run 1.35 to 1.55 times base pay. Computed as crew size times loaded rate divided by units per hour, so output misses inflate it invisibly.

Energy Cost per Unit

  • The electricity, gas, and compressed air consumed to make one unit. Industrial electricity averages 7 to 12 cents per kWh in the US, and energy runs 1 to 5 percent of cost for assembly but 15 to 40 percent for melting, heat treat, and drying processes. Submetering the big loads is the first step to managing it.

Gross Margin

  • Selling price minus direct cost, expressed as a percent of price. Job shops typically target 25 to 40 percent gross margin; high volume commodity work can run under 15. Margin and markup differ: a 25 percent markup on cost is a 20 percent margin on price, and mixing them up loses real money on quotes.

Capacity Utilization

  • The share of available machine or labor hours actually used for production. Healthy plants run constraint assets at 80 to 90 percent; past roughly 85 percent, queues and lead times grow nonlinearly. Utilization below 60 percent on a paid asset means the depreciation is being spread over too few parts.

Compliance Cost

  • The recurring cost of meeting regulatory and customer requirements: audits, certifications, testing, documentation, and training. ISO 9001 registration for a small plant runs several thousand dollars a year in audit fees alone, and regulated industries commonly spend 2 to 5 percent of revenue on quality and compliance overhead.

Packaging Cost

  • The material and labor to protect and ship product: cartons, dunnage, stretch wrap, labels, and pack labor. It typically runs 1 to 5 percent of product cost for durables and up to 10 percent or more for consumer goods. Cube utilization drives freight, so packaging design is also a freight decision.

Assembly Time

  • The hands-on time to join components into a finished unit, set by work content and layout rather than machine speed. Predetermined time systems price most manual actions in the 1 to 5 second range; part count and fastener count are the biggest levers, which is why design for assembly targets both first.

Operational Risk

  • The probability-weighted cost of things going wrong: equipment failure, quality escapes, supply interruptions, and safety events. Quantified as likelihood times consequence, often via FMEA scoring. A 2 percent chance of a 500,000 dollar line-down event is a 10,000 dollar expected cost worth insuring against with spares or redundancy.

Warranty Cost

  • The accrued cost of repairing or replacing product that fails in service. Durable goods makers typically reserve 0.5 to 3 percent of revenue for warranty; field failures cost 5 to 20 times the factory repair cost once freight, teardown, and goodwill are counted. Claim rate per thousand units is the tracking metric.

Quoting

  • Pricing a job from routing, material, and overhead data before winning it. Estimating accuracy within 5 percent of actuals is a strong shop; quotes built from stale routings commonly miss by 15 to 30 percent. Win rate near 20 to 35 percent usually signals healthy pricing; much higher often means quoting too low.

Rework

  • Correcting a nonconforming unit so it can ship, as opposed to scrapping it. Rework consumes capacity twice: once to build wrong and again to fix. A hidden factory of 5 to 15 percent rework is common in assembly plants, which is why first pass yield, not final yield, exposes the real cost.

Inspection

  • Verifying product against specification at receiving, in process, or before ship. Inspection detects rather than prevents, so its cost rises with defect rates upstream. Sampling plans trade risk for effort: a typical AQL based plan inspects 32 to 125 pieces per lot instead of 100 percent, at a defined consumer risk.

Yield Loss

  • Input that fails to become sellable output: scrap, trim, kerf, setup pieces, and rejects. Every point of yield on a material-heavy process is nearly a point of margin. Track loss by cause, because nesting loss, defect fallout, and startup loss respond to completely different fixes.

Test Cost

  • The capital and time to functionally verify product: test stands, fixtures, cycle time, and failure diagnosis. End of line test commonly runs 3 to 10 percent of unit cost in electronics and electromechanical assembly. Test stand utilization above roughly 85 percent makes test the hidden bottleneck of the plant.

Thermal Processing

  • Operations that heat material to transform it: heat treat, drying, curing, brazing, and melting. Energy dominates the economics, with furnace efficiency typically 40 to 80 percent depending on recuperation. Load density is the free lever: a half-loaded furnace burns nearly the same gas as a full one.

Material Cost

  • The purchased inputs consumed per unit, including yield loss and consumables, usually the largest single cost element at 40 to 70 percent of unit cost. Priced correctly it includes freight in, scrap allowance, and regrind or remnant credits. Indexing quotes to commodity prices protects margin on volatile inputs.

Field Service

  • Maintaining and repairing equipment at the customer site. A loaded service visit typically costs 300 to 1,500 dollars once travel, truck stock, and admin are counted, so first time fix rate, healthy at 75 to 85 percent, is the metric that controls cost per resolution.

Supplier Performance

  • How reliably a vendor delivers to specification, on time, at quoted cost. Scored on on-time delivery, healthy at 95 percent plus, incoming PPM defects, and responsiveness. A cheap part from a supplier at 90 percent OTD usually costs more than the price difference in expediting and line disruption.

Payback Period

  • The time for an investment's savings to repay its cost: investment divided by annual savings. Plants commonly require under 2 to 3 years for equipment and under 12 to 18 months for tooling and automation cells. Payback ignores savings after the cutoff, so pair it with a simple ROI percent for bigger bets.

Overhead Rate

  • Indirect plant cost allocated to products, usually as dollars per machine or labor hour. Machine rates typically run 40 to 150 dollars per hour depending on the asset. A stale overhead rate quietly misprices every quote, so recalibrate annually against actual spend and actual hours.

Water Usage

  • Process water consumed per unit or per day in washing, rinsing, cooling, and surface treatment. Industrial water plus sewer commonly costs 5 to 15 dollars per thousand gallons, and discharge permits cap flow and contaminants, so counterflow rinsing and recirculation typically cut consumption 30 to 70 percent.

Equipment Sizing

  • Choosing how much capacity to buy: machines, chambers, tanks, or vehicles, from demand, cycle time, and availability. Sized from peak sustained demand plus a 15 to 25 percent buffer, not average demand, because queues explode as utilization approaches 100 percent.

Electronics Assembly

  • Building circuit assemblies: solder paste, component placement, reflow, and test. Placement is priced per component, commonly 0.3 to 3 cents each at volume, and first pass yield at test, healthy above 95 percent, is the cost driver because diagnosing a failed board costs far more than placing it.

Production Scheduling

  • Deciding what runs on which resource and when, against due dates and capacity. Schedule adherence, healthy above 90 percent, measures whether the plan is real. Finite scheduling respects capacity limits; infinite loading reveals overloads. The schedule is only as good as the routing times underneath it.

Production Rate

  • Units produced per hour or per shift at a process, the denominator of every per-unit cost. Demonstrated rate over the trailing 90 days, not the routing standard, is the honest planning number; the gap between the two is typically 10 to 20 percent and represents recoverable capacity.

Inventory Turns

  • Annual cost of goods sold divided by average inventory value: how many times a year the stock converts to sales. Discrete manufacturers commonly run 4 to 12 turns; world class exceeds 20. Each added turn frees cash equal to one turn's worth of average inventory.

Digital Twin

  • A software model of a machine, line, or plant kept in sync with real operating data, used to test changes before touching the floor. Payback comes from avoided trials: simulating a layout or schedule change costs hours, while a failed physical trial costs days of output.

Manufacturing Analytics

  • Turning machine and quality data into decisions: dashboards, alerts, and models on top of downtime, cycle, and defect streams. The common failure is collection without action; a metric that no meeting reviews changes nothing. Start from the decision, then instrument the data it needs.

Line Speed

  • The linear rate of a continuous process in feet or meters per minute, set by the slowest stage in the train. Running above the capable speed trades scrap for output: a 5 percent speed increase that adds 2 points of scrap usually loses money on material-heavy products.

Process Efficiency

  • Actual output as a share of theoretical output for a machine, line, or energy system. The measurement window matters: an 85 percent efficient shift can hide a 60 percent efficient constraint. Always name the baseline, nameplate, demonstrated best, or standard, when quoting an efficiency number.

Part Weight

  • The mass of one unit, the multiplier that converts material price into material cost. In molding and extrusion, weight control IS cost control: running 3 percent over nominal weight on a resin part is 3 percent material cost given away on every unit, invisible without routine part weighing.

Measurement Accuracy

  • How close a measured value sits to the true value, bounded by instrument resolution, calibration, and technique. The 10 to 1 rule holds the instrument to a tenth of the tolerance being checked. Measurement error consumes tolerance: a gauge eating 30 percent of the band effectively tightens every spec.

Demand Forecasting

  • Estimating future orders to plan capacity, material, and staffing. Forecast error, commonly 20 to 40 percent at item level, is why safety stock exists; measuring it (MAPE by item family) tells you whether to fix the forecast or buffer against it. Aggregate forecasts are always more accurate than item ones.

Coating Coverage

  • The area one unit of coating covers at a specified film thickness. Theoretical coverage assumes 100 percent transfer; real processes deliver 30 to 70 percent depending on method, so effective coverage equals theoretical times transfer efficiency, and the gap is bought paint that never reaches the part.

Manufacturing Process

  • The sequence of operations that transforms material into product: forming, machining, joining, finishing, and assembly, each with its own cost drivers, capability limits, and quality risks. Process selection sets 70 to 80 percent of a part's lifetime cost before the first unit is made.

Workforce Training

  • Building and maintaining operator skills: onboarding, cross-training, and recertification. A new operator typically runs 60 to 75 percent of standard output for 4 to 8 weeks, so training cost is mostly lost production, not instructor time. A skills matrix per line exposes single-person dependencies before they become downtime.

Last reviewed 2026-05-12.