Quality Function Deployment (QFD) and House of Quality

Quality function deployment (QFD) is the structured method that takes what customers want and translates it, room by room, into the technical specifications your engineers and operations teams can actually build to. It closes the gap between a customer saying "I want it to last longer" and an engineer knowing exactly which material strength rating to target. Without that translation, product teams optimize for specs that feel important internally but don't connect to the purchase drivers that actually matter.

QFD was developed by Yoji Akao in Japan in the late 1960s and first deployed at the Kobe Shipyard division of Mitsubishi Heavy Industries in 1972. Toyota adopted it shortly after, reducing startup costs on new model launches by more than 60% within a decade. The method spread to U.S. manufacturers through the 1980s, carried largely by Ford and Xerox.

What Is Quality Function Deployment?

Quality function deployment (QFD) is a structured planning methodology that converts customer requirements (called "the whats") into specific technical characteristics (called "the hows") at each stage of product or service development. The output of each phase becomes the input for the next, so customer intent survives from concept through to final production.

The centerpiece of QFD is the House of Quality (HoQ), a matrix that visually maps customer needs against technical design parameters. It's called a "house" because the diagram includes a triangular roof that shows the correlations between technical characteristics themselves, sitting atop the main rectangular matrix like a roof on a building.

QFD connects tightly to Voice of the Customer (VoC) research. VoC collects raw customer language. QFD turns that language into actionable engineering targets. It also feeds directly from CTQ trees, which break broad customer needs into measurable quality characteristics. Those CTQs become the row inputs on the left side of the House of Quality.

Key Facts

• Toyota reduced pre-production engineering changes by 61% and startup costs by 61% after implementing QFD in the early 1980s (Hauser & Clausing, Harvard Business Review, 1988).
• Ford reported a 40% reduction in early production problems on vehicles designed using QFD compared to vehicles designed without it (American Supplier Institute, 1987).
• A study of 35 companies using QFD found that product development cycle time dropped by an average of 33% (GOAL/QPC research consortium, Quality Progress).

The House of Quality Explained

The House of Quality contains six distinct "rooms," each carrying a specific type of information. Understanding what goes in each room tells you exactly how to build one.

The relationship matrix is where most of the analytical work happens. Multiply each relationship score by the importance rating of the corresponding customer need, then sum down each column. The column with the highest score is the technical characteristic with the most impact on customer satisfaction. That's where engineering focus should go first.

The Four Phases of QFD

Most QFD deployments use a four-phase cascade. The output from Phase 1 feeds the input of Phase 2, and so on, so customer intent propagates all the way to the production floor.

Most teams complete Phase 1 (the classic House of Quality) and stop there. That's a useful starting point, but the real power comes from cascading all four phases, because it ensures that what the assembly-line operator does on Thursday morning traces directly back to a requirement a customer stated in a focus group six months earlier.

Common Mistakes

Listing features instead of needs. Customer needs should be expressed as outcomes ("opens smoothly in cold weather"), not product features ("use premium grease"). Features belong in the technical characteristics columns.

Skipping the competitive assessment. Teams that omit benchmarking often invest in technical improvements in areas where they're already ahead of the competition. The competitive column reveals where closing a gap actually changes buying behavior.

Treating all relationships as strong. When every cell in the relationship matrix gets a "9," the prioritization math collapses. Be disciplined: most cells should be empty or weak.

Building the roof but ignoring it. The correlation roof showing conflicts between technical characteristics is frequently constructed and then never acted on. Negative correlations (improving Spec A degrades Spec B) need explicit design decisions, not silence.

One-and-done. A House of Quality is a living document. Customer priorities shift. Competitor products change. Update the importance ratings and competitive benchmarks at least once per product generation.

How to Build a House of Quality

Step 1: Gather customer needs

Run Voice of the Customer research: interviews, surveys, focus groups, support ticket analysis. Capture verbatim language. Group related statements into themes and synthesize each theme into a clear need statement. Aim for 10-30 need statements. Too few and you miss nuance; too many and the matrix becomes unmanageable.

Step 2: Rate need importance

Ask customers (or use proxy data from purchase behavior, complaint frequency, or market research) to rate how important each need is on a 1-5 or 1-10 scale. These become your importance weights. They drive every downstream prioritization decision, so spend time on them.

Step 3: Define technical characteristics

List the engineering or process parameters your team can actually control: dimensions, tolerances, materials, cycle times, process temperatures, software response times. Each characteristic should be measurable and have a defined direction of improvement (more is better, less is better, or target is best).

Step 4: Build the relationship matrix

Work through each cell: does this technical characteristic affect this customer need? If yes, assign a score: strong (9), moderate (3), or weak (1). Leave the cell blank if there's no meaningful relationship. This step requires cross-functional input. Engineers, product managers, and quality leads should populate the matrix together.

Step 5: Score and prioritize

For each technical characteristic column, multiply every relationship score by the importance rating of the corresponding customer need, then sum the column. The result is the relative importance of that technical characteristic to overall customer satisfaction. Normalize the scores to percentages for easier comparison.

Step 6: Complete the correlation roof

For each pair of technical characteristics, assess whether improving one positively impacts the other (mark with a "+" or double "+"), conflicts with it ("-" or double "-"), or has no relationship (leave blank). Negative correlations need explicit resolution in the design phase.

Step 7: Run competitive benchmarking

Rate your current product and your top two or three competitors on each customer need, using a 1-5 scale. Plot these ratings in the competitive assessment room. Gaps between where you stand and where competitors stand (especially on high-importance needs) become your highest-priority design targets.

Step 8: Set technical targets

For each technical characteristic, set a specific, measurable target value. These come from the prioritization scores, competitive benchmarking gaps, and engineering feasibility. Targets become the outputs of Phase 1 and the inputs to Phase 2.

QFD Example

Here's a simplified House of Quality for a reusable water bottle, showing five customer needs mapped against four technical characteristics.

Reading the weighted scores: cap seal force (225) has the strongest link to overall customer satisfaction, followed by wall thickness (189), surface texture (141), and bottle weight (111). Engineering resources should prioritize cap seal force specifications first. This directly addresses the two highest-importance needs (doesn't leak, easy to open one-handed) at the same time.

Best Practices

Start small, then expand. Build your first House of Quality for a single product line or a single customer segment. A focused first run is far more useful than a sprawling matrix that never gets finished.

Cross-functional team from day one. QFD fails when only the quality team fills out the matrix. You need voice-of-customer input (product management or customer success), technical expertise (engineering or operations), and strategic context (product leadership or sales) in the same room.

Link QFD to DMADV and Six Sigma projects. QFD Phase 1 outputs map directly to the Define and Measure phases of DMADV. When you're designing a new product or process from scratch, QFD is how you ensure the CTQs you're optimizing for are the ones customers actually care about.

Use the Kano Model to sort needs first. Before weighting customer needs by importance, run a Kano analysis to separate basic needs (expected, won't create delight if present), performance needs (more is better), and excitement needs (unexpected, creates delight). Basic needs get high importance ratings by default; excitement needs may get moderate importance but high strategic weight.

Revisit after competitive moves. A competitor launching a better product changes the competitive benchmarking room. Reassess importance ratings and targets whenever the market shifts. Total Quality Management programs use QFD as a recurring planning tool, not a one-time exercise.

Connect to Value Stream Mapping. After Phase 3 (process planning), the process parameters you've identified are exactly the inputs your value stream map should measure. This creates a direct link from customer requirement to process step to value-stream waste analysis.

Frequently Asked Questions

What is the difference between QFD and the House of Quality?

QFD is the full four-phase methodology. The House of Quality is the specific matrix used in Phase 1 of QFD. People often use the terms interchangeably because Phase 1 is the most commonly implemented part of the framework, but the House of Quality is one tool within the broader QFD system.

How long does it take to build a House of Quality?

A focused cross-functional team with good VoC data can complete Phase 1 in two to four half-day workshops. Larger products with 20+ customer needs and 15+ technical characteristics take longer. The common mistake is trying to complete it solo or in a single marathon session. Spread it across sessions with preparation work in between.

When should you use QFD instead of just a requirements document?

Use QFD when you need to make explicit trade-off decisions across competing technical characteristics, when you want to ensure engineering priorities trace back to customer data (not internal assumptions), or when you're entering a competitive market where benchmarking against alternatives matters. A requirements document tells you what to build; QFD tells you why those requirements are prioritized the way they are.

Can QFD be used for services, not just products?

Yes. Service QFD replaces physical dimensions with service attributes: wait time, interaction script quality, system uptime, resolution rate. The matrix structure is identical. Many banks, hospitals, and logistics companies use QFD to design service processes.

How does QFD connect to CTQ trees?

A CTQ tree is often the prep work for a House of Quality. The CTQ tree takes a broad customer need and breaks it into measurable quality characteristics. Those measurable characteristics become the row inputs (customer needs) on the left side of the House of Quality matrix. Together they form a clean pipeline: VoC research feeds the CTQ tree, which feeds the House of Quality, which feeds the engineering spec.

---

QFD won't tell you what your customers want. That's what Voice of the Customer research is for. But once you know what they want, QFD gives you the structure to turn that knowledge into technical decisions that actually get built. Teams that skip this step often build technically excellent products that still disappoint customers, because the excellence was aimed at the wrong targets.