In a factory, quality means that products meet their specifications and do what customers expect of them. It isn’t a single number. It’s several measurements of different things: how many parts come out right, how consistently a machine hits its target, and what mistakes cost. The biggest shift in modern manufacturing is treating quality as something built into the process rather than something sorted out by inspecting at the end.1Exact Metrology. Quality Assurance vs. Quality Control: Why Inspection Alone Does Not Create Good Parts
The Different Things Quality Can Mean
Quality is several things at once. One widely used framework breaks product quality into eight parts: performance, features, reliability, conformance, durability, serviceability, aesthetics, and perceived quality.2International Journal of Quality & Reliability Management. How product quality dimensions relate to defining quality A car can score high on features yet low on reliability, so “good quality” depends on which part you mean.
For a washing machine, performance is how clean the clothes get, and features are the extra cycles. Reliability is how rarely it breaks in the first years, while durability is how many years it lasts overall. Serviceability is how easily someone can repair it, aesthetics is how it looks, and perceived quality is the brand reputation that shapes what buyers assume before they’ve used it.
Quality is also not the same as luxury. A cheap plastic bucket can be high quality if every one meets its spec, and an expensive product that varies from unit to unit is poor manufacturing quality. Factories judge quality against the design, not against a fancier alternative.
Of the eight, conformance is what factories measure most. Each part has a tolerance, which is an allowed range around its target size, such as a hole of 10 mm plus or minus 0.05 mm. A part inside the allowed range conforms, and one outside it is a defect.
High conformance doesn’t guarantee happy customers, though. A part can match its drawing perfectly and still disappoint if the drawing was wrong or missed what buyers wanted. Conformance tells you whether the factory made what the design asked for, not whether the design was the right one.
Quality Assurance Versus Quality Control
Quality assurance (QA) happens before and during production and focuses on the system, aiming to stop defects from happening. Quality control (QC) happens after production or after a step, focuses on the result, and catches defects before they reach the customer.3Fabrico. Quality Assurance vs. Quality Control in Manufacturing: What is the Difference?
On a shop floor, prevention looks like training operators, maintaining machines, and writing clear work instructions before a run. Checking finished work looks like an inspector measuring a sample of parts and rejecting the ones that fail. Both are needed: the checking confirms the prevention is working and catches what slips through.
Inspection alone falls short. In much of manufacturing, parts are still made first, measured afterward, and compared against their allowed range.1Exact Metrology. Quality Assurance vs. Quality Control: Why Inspection Alone Does Not Create Good Parts By the time a size drift is found, more bad parts may already have been made. Inspection then becomes sorting good from bad rather than controlling the process.1Exact Metrology. Quality Assurance vs. Quality Control: Why Inspection Alone Does Not Create Good Parts1Exact Metrology. Quality Assurance vs. Quality Control: Why Inspection Alone Does Not Create Good Parts
The Key Numbers Factories Use to Measure Quality
- Quality rate: good parts made without defects or rework, divided by total parts produced.4MPDV. Overall Equipment Effectiveness / OEE
- Defects per million opportunities (DPMO): how many defects occur for every million chances to make one.
- Sigma level: a score built from DPMO, where higher means fewer defects.
- Cpk: a capability score showing whether a process stays comfortably inside its allowed range; 1.33 or higher is the usual target.5Sensors and Materials. Implementing Process Capability Index (Cpk) for Effective Product Quality Stabilization: The Case of a Lock Manufacturing Company
The quality rate is also one of three parts of Overall Equipment Effectiveness, a common score for how productively a machine runs.4MPDV. Overall Equipment Effectiveness / OEE A machine running fast but making 5% scrap loses points on quality even if it never stops.
Reading these numbers takes some care. A 98% quality rate sounds high, but on 100,000 parts it means 2,000 bad ones. Six Sigma means about 3.4 defects per million, the benchmark behind the name. And it’s more useful to compare numbers over time for the same process than across different products.
Factories track measurements over time on control charts, which flag unusual swings so their causes can be found and fixed.6Materials and Design. Statistical process control in machining, a case study for machine tool capability and process capability In one lock factory, workers measured a critical dimension twice a day for 20 days and found a Cpk of 0.72, well below the 1.33 target.5Sensors and Materials. Implementing Process Capability Index (Cpk) for Effective Product Quality Stabilization: The Case of a Lock Manufacturing Company Engineers fine-tuned the lathe’s tool position to centre the dimension on its 3.51 mm target. A second 20-day check gave a Cpk of 1.54, meaning the process was now comfortably capable and unlikely to produce out-of-spec parts.5Sensors and Materials. Implementing Process Capability Index (Cpk) for Effective Product Quality Stabilization: The Case of a Lock Manufacturing Company
The capability score beats a simple average because a process can average right on target yet swing wildly and make many bad parts. It captures both how centred the process is and how much it varies. Below 1, some parts will fall outside the allowed range. A score based on too few readings, or on readings from one unusually calm stretch, can look better than the process really is.
Sigma level shows progress in defect counts. A rubber glove maker that cut leaking gloves by about half brought DPMO from about 195,000 to about 84,000.7International Journal of Lean Six Sigma. A Six Sigma and DMAIC application for the reduction of defects in a rubber gloves manufacturing process Its sigma level rose from 2.4 to 2.9, a real gain but still far from the 6 that Six Sigma aims for.
How Parts Are Physically Checked
For size and shape, touch-probe measuring machines and CNC machine tools are a basic method because they’re highly accurate, robust, and work on almost anything.8International Journal of Extreme Manufacturing. Surface form inspection with contact coordinate measurement: a review Newer machines combine touch probes with optical and laser sensors to check many shapes faster.9International Journal of Precision Engineering and Manufacturing. A Comprehensive Review of Smart Coordinate Measuring Machines: Architectural Innovations, Multi-Sensor Integration, and Uncertainty Evaluation for Industry 4.0
Cameras watch for visible defects. Many plants place automatic camera inspection at several stations between critical steps, as Bosch does in car electronics.10Procedia CIRP. Design and development of an Automatic Optical Inspection (AOI) system support based on digital manufacturing Automated inspection doesn’t tire or vary the way human inspectors do, and it keeps high precision over long shifts.11Expert Systems with Applications. Computer vision defect detection on unseen backgrounds for manufacturing inspection One system for woven fabric spotted and sorted eight defect types, such as stains, holes, and broken threads, at 20 metres per minute with about 97% accuracy.12PubMed Central. Automated Optical Inspection for Defect Identification and Classification in Actual Woven Fabric Production Lines
Some flaws are hidden. Non-destructive testing (NDT) checks parts without damaging them, using methods like visual checks, dye penetrants, magnetic particles, eddy currents, ultrasound, and X-ray.13Virtual and Physical Prototyping. Applications of non-destructive testing techniques for post-process control of additively manufactured parts Ultrasound and X-ray can find typical hidden flaws even in 3D-printed metal parts.14Additive Manufacturing. Non-destructive testing application of radiography and ultrasound for wire and arc additive manufacturing For critical bonded aerospace joints, combining ultrasound and X-ray data can give a more reliable verdict.15Applied Sciences. Non-Destructive Evaluation of the Quality of Adhesive Joints Using Ultrasound, X-ray, and Feature-Based Data Fusion
People still have a place. They handle unusual defects the cameras weren’t trained on, and they set up, verify, and audit the automated systems. Human checks are least reliable on long, repetitive runs, where fatigue sets in.
Even the same measurement can happen in different places. A separate climate-controlled lab is very accurate but slow, so parts wait for results. Measuring at the machine gives faster feedback, so an operator can adjust before many more parts are made.
What Poor Quality Costs
- Prevention: training, maintenance, and process engineering that stop defects.16Fabrico. COPQ: Cost of Poor Quality Formula, Worked Example, 4 Categories
- Appraisal: inspection labour, testing, and audits that detect defects.16Fabrico. COPQ: Cost of Poor Quality Formula, Worked Example, 4 Categories
- Internal failure: scrap, rework, re-inspection, and downtime caught before shipping.16Fabrico. COPQ: Cost of Poor Quality Formula, Worked Example, 4 Categories17Autodesk. What Is Quality in Manufacturing and How Is It Measured?
- External failure: warranty repairs, returns, recalls, penalties, and lost sales once a defect reaches the customer.16Fabrico. COPQ: Cost of Poor Quality Formula, Worked Example, 4 Categories17Autodesk. What Is Quality in Manufacturing and How Is It Measured?
These costs can be traced to specific inputs like materials and machines as well as the whole company.18International Journal of Quality & Reliability Management. The relationship between quality and quality cost for a manufacturing company Spending more on prevention usually lowers the more expensive failure costs.
One defect costs very different amounts depending on when it’s caught. At the design stage, a flawed part costs a drawing change. On the line, it costs scrap and rework time. Once the customer finds it, the bill includes shipping, repairs, and reputation, and a recall can multiply that across every unit sold.
How Factories Find and Fix Root Causes
The eight disciplines (8D) method walks a team through forming, describing the problem, containing it, finding and verifying root causes, choosing and putting fixes in place, preventing recurrence, and recognising the team.19Applied Sciences. Improving a Manufacturing Process Using the 8Ds Method. A Case Study in a Manufacturing Company Teams often map possible causes on a fishbone diagram alongside it.20Materials Today: Proceedings. Investigation and implementation of 8D methodology in a manufacturing system
The early steps set up everything after. The team is built from people who see different parts of the problem, such as an operator, a maintenance technician, and a quality engineer. The problem description covers what is wrong, where it appears, when it started, and how many parts are affected. A vague description like “parts are bad” sends the team looking in the wrong place.
Containment comes before the fix because the permanent fix can take weeks of analysis. Sorting stock, for example, stops bad parts reaching customers right away. Skip it, and customers keep receiving defects in the meantime.
A real root cause is often not where people look first. An auto-parts supplier tracing warped aluminium pins found the cause was workers handling hot material by hand. An automation failure had forced manual handling, made worse by no standard instructions or training.21Eng. Application of Statistical Quality-Analysis Tools in the 8D Methodology for Manufacturing Quality Improvement The fix was in the process and training, not just the moulding machines.
Six Sigma has its own route, DMAIC (define, measure, analyse, improve, control), a step-by-step way of finding and removing the causes of defects.22Management Science Letters. Defect reduction in a capacitor manufacturing process through Six Sigma concept: A case study It’s how the glove maker cut leaking gloves by half, by tuning two process settings.7International Journal of Lean Six Sigma. A Six Sigma and DMAIC application for the reduction of defects in a rubber gloves manufacturing process
Preventing Mistakes Before They Happen
Poka-yoke, Japanese for mistake-proofing, designs a step so it can’t be done wrong. Examples include a fixture that won’t clamp a part loaded backwards and a sensor that stops a press if a part is missing. On one starter assembly line, poka-yoke forcing correct order and full pressing of two parts led to zero defects in a 500-piece check.23International Journal of Business and Systems Research. Implementation of poka-yoke to achieve zero defects in an assembly line of a limited company
It beats reminders because warning signs and checklists depend on a person remembering, and attention drops on repetitive work. Poka-yoke makes the wrong action physically impossible or immediately obvious. It is often cheap, such as a notch or guide pin, which is why it is usually tried before costly inspection equipment.
Factories also increasingly use sensors in place of visual checks, and machine learning to predict which parts will fail.24Processes. A Review on Data-Driven Quality Prediction in the Production Process with Machine Learning for Industry 4.0 At one car plant, a prediction model cut the cars needing a check from 1,530 to 202.25PubMed Central. A Predictive Quality Inspection Framework for the Manufacturing Process in the Context of Industry 4.0 It caught only about 59% of faulty cars, so prediction narrows inspection rather than replacing it.25PubMed Central. A Predictive Quality Inspection Framework for the Manufacturing Process in the Context of Industry 4.0 Similar models can predict machining quality and support maintenance decisions.26Journal of Intelligent Manufacturing. From knowledge-based to big data analytic model: a novel IoT and machine learning based decision support system for predictive maintenance in Industry 4.0
What Quality Certifications Like ISO 9001 Mean
ISO 9001 is part of the ISO 9000 family of standards, which guide companies in meeting customer requirements and continually improving quality.27Production Engineering Archives. Recertification of a Quality Management System based on ISO 9001 – is it a must for a modern manufacturing company? The 2015 version added a requirement to identify and manage risks.28Total Quality Management & Business Excellence. Assessment of ISO 9001: 2015 implementation: focus on risk management approach requirements compliance in an automotive company Certification shows a managed system, not that every product is flawless.
Getting certified starts with a company writing down how it runs key processes, such as handling customer orders, controlling documents, and dealing with defects. An outside certification body then audits the company to check that it actually follows what it wrote. Follow-up audits review certification periodically, so a buyer should check that it is still valid.
Asking for real metrics like defect rates or Cpk adds detail the certificate doesn’t give. Stricter industry versions exist for cars and aerospace.
In car manufacturing, the Production Part Approval Process (PPAP) requires suppliers to prove their process can make parts meeting all customer requirements during real production.29International Journal of Engineering Research in Africa. Role of Production Part Approval Process in Continuous Quality Improvement and Customer Satisfaction It’s tied to advance quality planning, which supplies the evidence PPAP asks for.30Archives of Materials Science and Engineering. Tools of product quality planning in the production part approval process
Quality records also matter legally. Product liability suits typically claim negligence, breach of warranty, or strict liability.31Taylor & Francis Online. Product Liability: Beyond Loss Control–An Argument for Quality Assurance Courts look at how a company’s quality system documented and handled problems, not just the defect.31Taylor & Francis Online. Product Liability: Beyond Loss Control–An Argument for Quality Assurance Good records help a company show due diligence.

