Product Quality: Garvin's 8 Dimensions, KPIs & How It's Measured

What is product quality?

Product quality is the degree to which a product conforms to specification and meets the expectations of the customer who uses it. That one-sentence definition hides two very different perspectives that quality professionals deliberately keep apart: conformance to specification (a manufacturer's view — does the part match the drawing?) and fitness for use (a customer's view — does it do what I wanted?). A product can be perfect on the first measure and mediocre on the second, which is why product quality is never a single number.

Before going further, one disambiguation worth getting right: product quality is not the same as process quality. Process quality describes how consistently and capably a production process runs. Product quality describes what comes out of it. Strong process quality (high Cpk, low drift) is usually the most reliable path to strong product quality — but the two are measured differently, owned by different people, and improved through different levers.

Garvin's eight dimensions — the canonical framework

David A. Garvin's 1987 Harvard Business Review article Competing on the Eight Dimensions of Quality is still the most complete framework for what "quality" actually means from a customer perspective. Most textbook summaries drop it to four or five dimensions and lose the analytical power in the process. The full eight:

Most production-floor quality work sits firmly in dimensions 3–5 — reliability, conformance, durability. These are the ones a manufacturer controls directly. Dimensions 1, 2 and 7 are set in design and marketing. Dimension 8 is won and lost over years.

The Kano model — a complementary lens

Garvin's framework describes what quality is. Noriaki Kano's model describes how each quality attribute affects customer satisfaction, which is a different and often more useful question. Kano splits attributes into five categories:

• Basic (must-be) — absence causes dissatisfaction; presence produces no delight. Brakes in a car. A usable power button. Customers take these for granted.
• Performance (one-dimensional) — satisfaction scales linearly with the attribute. Battery life, fuel economy, cycle time.
• Excitement (attractive) — presence delights; absence is unnoticed. Features customers didn't know to ask for.
• Indifferent — no impact on satisfaction either way.
• Reverse — more of the attribute actively dissatisfies some segments.

The Kano insight relevant to manufacturing: conformance defects almost always hit the Basic category. A loose screw, an out-of-tolerance dimension, an assembly error — these rarely disappoint in a measurable way (because customers assume correctness) but they generate outsized dissatisfaction when they occur. That asymmetry is the reason conformance quality gets disproportionate attention in regulated and safety-critical industries.

How product quality is actually measured

Philosophical frameworks are useful. What a quality manager needs on a Monday morning is numbers. The standard set of product-quality KPIs, in roughly the order they appear in most factories:

Two observations from reviewing these KPIs in actual factories. First, most mid-market plants track FPY and scrap rate well, but struggle with RTY because it requires clean data across every step — which requires an MES-level data layer. Second, Cpk is the single KPI where the gap between "we measure it" and "we use it to steer the process in real time" is widest. Calculating Cpk from weekly samples is normal. Calculating it live and alarming on drift is where modern quality programmes separate from legacy ones.

The standards layer — ISO 9001 and IATF 16949

For most mid-market manufacturers, two standards define the formal framework around product quality. ISO 9001 (general quality management systems, currently the 2015 edition with a 2026 revision in progress) establishes the baseline: documented processes, management responsibility, continuous improvement, risk-based thinking. It applies across industries. IATF 16949 is the automotive-specific extension — much stricter on traceability, PPAP, APQP, FMEA, and escape prevention. If you supply the automotive industry, IATF 16949 certification is effectively non-optional, and the audit burden it creates drives a significant share of quality-system investment.

Both standards share a DNA: they require that quality claims be evidenced, not just asserted. That requirement is what drives most MES and SPC investment in quality-critical industries. You cannot audit what you have not measured.

From process data to product quality — the SYMESTIC angle

Product quality emerges from thousands of process decisions per day — a temperature setpoint held or missed, a cycle completed within tolerance or just outside it, a tool change performed on schedule or one cycle late. Each decision is usually invisible by the time the finished product ships. The link between the two is process data: captured continuously at the machine, tied to a specific production order and serial number, and correlated with any reject event that occurs. That correlation is what turns a quality programme from reactive (we caught defects at the end) to preventive (we caught drift before the first defect). Across the 15,000+ machines SYMESTIC has connected, this pattern repeats consistently — the data needed to improve product quality usually already exists at the PLC level. The work is getting it into a system where it can be used.

FAQ

What is the difference between product quality and service quality?
Product quality applies to tangible goods and is primarily measured against specifications and physical performance. Service quality applies to intangible deliverables (banking, consulting, hospitality) and is measured along different dimensions — responsiveness, empathy, reliability of delivery, assurance. Parasuraman's SERVQUAL model is the service-side equivalent of Garvin's eight dimensions.

Are Garvin's eight dimensions still relevant in 2026?
Yes — the framework is nearly forty years old and still holds up because it describes quality from the customer's perspective, which hasn't changed. What has changed is how the underlying data is captured: real-time sensor data, automated inspection, digital traceability. The dimensions are the same; the instrumentation is vastly better.

Is product quality the same as zero defects?
No. Zero defects refers specifically to conformance quality (Garvin's dimension 4). A product can have zero defects and still be low-quality if it performs poorly, lacks reliability, or looks and feels cheap. Conformance is necessary but not sufficient.

What's the single most useful product-quality KPI to start with?
First Pass Yield, broken out by machine and by shift. It is simple to calculate, directly actionable on the shopfloor, and leads naturally to the next KPIs (Cpk when you want to understand why FPY varies, RTY when you want to understand the end-to-end picture). Starting with warranty returns is a common mistake — the signal is too delayed and aggregated to drive daily improvement.

Related: Production Defect · Statistical Process Control · Cp/Cpk · First Pass Yield · OEE · ISO 9001

About the author

Christian Fieg

Head of Sales at SYMESTIC. 25+ years in manufacturing, including three years as a Six Sigma Black Belt in automotive Headliner production. Former global MES & traceability lead at Johnson Controls. IATF-16949 environments across four continents. Author of "OEE: One Number, Many Lies" (2025). · LinkedIn