A factory hires more inspectors to catch the faulty units before they ship. Defects keep slipping through, returns climb, and the inspectors get blamed. The mistake is older than the factory: treating quality as a thing you check for at the end, rather than something you design into the work from the start. Quality management is the body of theory and tools that fixes the order, quality first, inspection second, and the process itself as the thing you actually manage.

The quick version

  • Quality management is the deliberate practice of making sure a product or service reliably meets customer needs, by managing the process that produces it, not just inspecting the output.
  • The big idea, from W. Edwards Deming: stop depending on mass inspection and build quality in. Inspection finds defects too late; you want a process that rarely makes them.
  • Joseph Juran framed the work as three jobs, plan, control, improve, and Philip Crosby showed that poor quality is expensive, so prevention usually pays for itself.
  • ISO 9001 is the international standard that turns these ideas into an auditable management system, built on seven principles starting with customer focus.

The idea in depth: stop inspecting, start building

The intellectual centre of quality management is one counter-intuitive move, and it belongs to W. Edwards Deming, the American statistician who helped rebuild Japanese manufacturing after the war. In his 14 Points for Management, set out in Out of the Crisis (MIT Press, 1986), point three reads: "Cease dependence on inspection to achieve quality. Eliminate the need for inspection on a mass basis by building quality into the product in the first place." Inspection, in this view, does not create quality, it only sorts the good from the bad after you have already spent the money making both.

That points your attention upstream. Instead of asking "how do we catch more defects?", ask "what in our process is producing them, and can we change it so it stops?" Deming's deeper claim, set out in his System of Profound Knowledge, and specifically its "knowledge of variation" component, is that most variation in output comes from the system the worker operates inside, not from the worker. Blaming people for a system fault is both unfair and useless; it leaves the actual cause untouched. The practical version for a leader: when results vary, interrogate the process before you appraise the person.

An honest limitation. Deming's most quoted figure, often rendered as "94% of problems belong to the system, not the worker", is his estimate, restated across his lectures, not a measured constant; the real split varies enormously by setting. Treat it as a corrective bias against scapegoating, not a precise law. The principle survives even if the number is soft: before you discipline a person for a bad result, check whether the system handed them a bad hand.

Three jobs, not one: Juran's trilogy

If Deming supplied the philosophy, Joseph Juran supplied the management structure. Juran, whose Quality Handbook remains a standard reference, now in its seventh edition, argued that managing quality is really three distinct jobs, which the field calls the Juran Trilogy: quality planning (work out who the customer is and design a process capable of meeting their needs), quality control (run the process and hold it to standard, acting on problems as they appear), and quality improvement (deliberately raise the standard by attacking the root causes of recurring defects).

The trilogy earns its keep because most organisations are unbalanced across the three. Plenty of teams have control, dashboards, checklists, someone watching the numbers, but almost no planning and no improvement, so they firefight the same defects forever. The useful question is which of the three you are actually doing this quarter. Holding the line (control) is not the same as raising it (improvement), and the latter needs a deliberate, resourced project; it does not happen as a by-product of watching a dashboard.

flowchart LR
  A(["Quality planning
design a capable process"]) --> B(["Quality control
hold it to standard"])
  B --> C(["Quality improvement
attack root causes, raise the bar"])
  C -.->|"new, higher standard"| B
Juran's trilogy: plan it, hold it, then deliberately raise it, and the new level becomes the next thing you control. Leaders Loop

An honest limitation. The trilogy is a map, not a method, it tells you the three jobs exist but not how to do them. Quality planning in a hospital, a software team and a steel mill look nothing alike, and Juran's own work leans heavily on manufacturing. Use the trilogy to diagnose what is missing; reach for domain-specific tools (statistical process control, root-cause analysis, design reviews) to actually do the work.

Why prevention pays: the cost of poor quality

The objection a finance-minded leader will raise is that all this prevention costs money the business does not have. Philip Crosby's answer, in Quality Is Free (1979), is that poor quality already costs you far more, you just are not measuring it. Crosby split quality costs into the price of conformance (preventing and checking) and the price of non-conformance (scrap, rework, returns, lost customers). His estimate was that non-conformance could run to around 20% of sales in manufacturing and higher in services, while doing things right in a well-run company cost a few percent. The framework most teams use to make this concrete is the cost-of-quality model: prevention and appraisal costs (spent to get it right) versus internal and external failure costs (paid when you get it wrong).

The leadership job here is to make the invisible cost visible. Most failure cost stays hidden because nobody totals the rework hours, the returned units and the discounts handed over to keep an angry customer, it is scattered across a dozen budgets. Put a rough number on it for one product line and the case for prevention usually makes itself: a defect caught at the design stage is far cheaper than the same defect caught by a customer, because by then you have paid to build it, ship it, support the complaint, and rebuild trust.

A defect doesn't get cheaper the longer it survives. It gets more expensive at every step toward the customer.

This is also where the international standard earns its keep. ISO 9001, the world's most widely used quality-management standard, does not tell you what your product should be; it asks you to run a disciplined system around it, resting on seven principles set out in ISO 9000:2015: customer focus, leadership, engagement of people, the process approach, improvement, evidence-based decision-making, and relationship management. The honest caveat: a certificate proves you have a documented system, not that your product is good. Plenty of certified organisations ship mediocre work with immaculate paperwork. Treat ISO 9001 as scaffolding for the discipline, not a substitute for it.

A worked example

Take a mid-sized contract manufacturer, call it Tamar Components, making metal brackets for appliance makers. (Illustrative figures throughout; this is a teaching example, not a real company.) Customers keep rejecting batches for out-of-tolerance holes. The reflex is to add a final inspection gate and hire two more checkers, at an illustrative £80,000 a year. Rejections fall, but only because more parts are now caught in-house and scrapped, the underlying defect rate has not moved, and scrap is climbing.

A quality manager runs the problem through the trilogy instead. Control data shows the holes drift out of tolerance late in each shift. Improvement, a short root-cause project, traces it to a drill bit that dulls predictably after about 4,000 holes, but is only changed when it visibly fails. The fix is upstream: change the bit on a schedule, not on failure. Planning bakes it in, so the next process is designed with the wear limit built into the maintenance standard.

flowchart TD
  A(["Problem: customers reject
out-of-tolerance batches"]) --> B{"Inspect harder,
or fix the process?"}
  B -->|"Add inspectors (£80k/yr)"| C(["Defects still made,
scrap rises, cost hidden"])
  B -->|"Root-cause the defect"| D(["Drill bit dulls after
~4,000 holes"])
  D --> E(["Change bit on schedule;
bake wear limit into the standard"])
  E --> F(["Fewer defects made at all
prevention beats appraisal"])
The same problem, two responses: appraisal hides the cost of the defect; prevention removes the cause. Leaders Loop

The defect rate falls because the parts are no longer made wrong, not because more are caught. That is Deming's point three in one shift of attention, Juran's improvement job in one project, and Crosby's arithmetic in one budget line: the cost of a planned bit-change is a rounding error against the scrap, rejects and inspection it removes. Reverse the order, inspect first, never ask why, and you pay forever for defects you keep manufacturing.

Frequently asked questions

Is quality management just for manufacturing?

No, it started there because factories made variation easy to measure, but the principles travel. Hospitals use them to reduce medication errors, software teams use them to cut defect-escape rates, and service businesses use them to make a customer's experience consistent. The vocabulary changes (a "defect" might be a billing error or a missed SLA), but the core move, manage the process, do not just inspect the output, is general.

What's the difference between quality assurance and quality control?

Quality control (QC) is checking the output against the standard, inspecting, testing, measuring after the fact. Quality assurance (QA) is the broader, upstream work of designing processes so they produce good output in the first place. In Juran's terms, QC is the "hold it to standard" job; QA spans planning and improvement. Both matter, but the field's whole argument is that leaning only on QC is the expensive way round.

Do we need ISO 9001 certification?

It depends on your market. Certification is genuinely useful when customers or regulators demand it, or when you need an external discipline to make quality habits stick. But it is a means, not the goal: a certificate documents that you follow your own processes, not that those processes are any good. If you are pursuing it, pursue the underlying discipline first and let the certificate confirm it, not the other way round.

Isn't "build quality in" just common sense?

It sounds obvious until you watch where organisations actually spend. The default reflex under pressure is to add an inspection step, write a new checklist, or hire a checker, all downstream, all appraisal. Genuinely building quality in means changing the process that makes the defect, which is slower, harder, and less visible than catching it. The idea is simple; resisting the easier, costlier habit is not.

How do I start without a big programme or budget?

Pick one recurring defect that annoys customers, and instead of adding a check for it, spend an hour asking why it happens, keep asking until you reach a process cause you can change. That single root-cause habit, repeated, is quality improvement. You do not need a certification project or a quality department to begin; you need to stop treating the symptom and fix one cause this week.

Related in the Toolkit

Quality management is one lens on how work flows through an organisation, so it sits close to the broader improvement disciplines, the same root-cause instinct powers Lean and Six Sigma, and you cannot improve a process you have not first mapped and understood.

Where to go next