Quality Function Deployment (QFD)

Quality Function Deployment (QFD) is a structured method for translating what customers want into specific engineering and production requirements, so design priorities are driven by the voice of the customer instead of internal guesswork. Its signature tool is the House of Quality, a matrix that connects customer wants to measurable design characteristics and shows which of those characteristics matter most. Done well, QFD stops a design team from optimizing the things that are easy to measure and ignoring the things the customer actually judges you on.

QFD is a planning discipline, not a form to fill in. The matrices are just where the thinking gets recorded; the value is the argument the team has while building them. This guide covers what QFD is, how to read and build a House of Quality, a worked example, how the method cascades from a customer wish list all the way to the plant floor, and the ways it commonly goes wrong. It sits alongside the design-side quality tools in the broader lean and quality toolkit.

What Is Quality Function Deployment?

Quality Function Deployment is a systematic way to carry customer requirements through every stage of development, from concept to design to process to production, without losing them along the way. The problem it solves is translation loss. A customer says "I want a door that feels solid"; engineering hears something vague and substitutes its own priorities; by the time the part reaches the plant, the original want has evaporated. QFD forces each translation to be explicit: this customer want maps to these measurable characteristics, weighted this heavily, with these targets. The voice of the customer that feeds it can come from interviews, warranty claims, complaint logs, or lost-sale reasons, but it has to come from customers, the discipline collapses the moment the team starts inventing what it thinks customers meant.

The method was developed by Yoji Akao in Japan in 1966, and its power was demonstrated at the Mitsubishi Heavy Industries Kobe shipyard, where the House of Quality matrix was applied to complex ship design around 1972. From there it spread to automotive and electronics and became a staple of structured product development worldwide. It pairs naturally with critical-to-quality thinking and feeds directly into design FMEA because once you know which characteristics matter most, you know where failure would hurt most.

What Is the House of Quality?

The House of Quality is the first and most famous QFD matrix, named for its shape: a grid of rooms with a triangular "roof" on top. Each room holds one part of the translation. The left wall lists the customer wants (the "whats") with an importance rating; the ceiling lists the engineering characteristics that could satisfy them (the "hows"); the central relationship matrix scores how strongly each how affects each what. The roof is a correlation triangle showing where the hows help or fight each other, and the basement holds the calculated priorities, targets, and a competitive comparison. Because every judgment sits in one room of a single picture, the House of Quality doubles as a piece of visual management for the design decision: anyone can see what drove a priority and challenge it, instead of the reasoning living in one engineer’s head.

How Do You Build a House of Quality?

You build it room by room, starting with the customer and ending with prioritized targets. The order matters, because each step feeds the next:

1. Capture the customer wants. Gather the voice of the customer and phrase each want in the customer's own terms, "the cap seals on the first turn," not "torque spec." Group and de-duplicate them into the left wall.
2. Rate the importance of each want. Have customers (not the team) weight them, commonly on a 1–5 scale. This weighting drives every priority the matrix later calculates, so it is worth getting from real customers.
3. List the engineering characteristics. For each want, ask "what measurable characteristic would tell us we delivered it?" These are the hows across the ceiling, each one measurable, each with a direction of improvement.
4. Score the relationship matrix. For every want-versus-characteristic cell, score the strength of the link, usually 9 (strong), 3 (moderate), 1 (weak), or blank. A row with no strong relationship means a customer want that nothing you plan to measure will satisfy, a red flag.
5. Fill in the roof correlations. Mark where characteristics reinforce each other and, more importantly, where they conflict. Conflicts are where the real engineering trade-offs live and where a target on one spec quietly hurts another.
6. Add competitive and technical benchmarks. Rate how you and competitors satisfy each want, and benchmark the current value of each characteristic. This grounds the targets in reality instead of ambition.
7. Calculate priorities and set targets. Multiply each importance by its relationship scores and sum down each column. The columns with the highest totals are the characteristics that most drive customer satisfaction, that is where the design effort and the tightest targets go.

What Does a Worked House of Quality Look Like?

Here is a stripped-down example for a squeeze-bottle closure. Three customer wants sit on the left with importance weights; three engineering characteristics run across the top; the cells hold relationship strengths (9/3/1). The bottom row multiplies and sums to reveal which characteristic matters most.

The arithmetic is simple: seal torque scores (5×9)+(4×9)+(3×9)=108, thread pitch (5×9)+(4×3)+(3×3)=66, wall thickness (5×1)+(4×3)+(3×1)=20. The matrix says seal torque is the characteristic that most drives customer satisfaction here, so it earns the tightest target, the most design attention, and the closest watch in production. (These numbers are an illustration, not data from any product.) The value is not the ranking itself; it is that the ranking is now visible, defensible, and tied explicitly to what the customer said.

How Does QFD Cascade Beyond the First Matrix?

Full QFD does not stop at one house. In the classic four-phase model, the outputs of each matrix become the inputs of the next, carrying the prioritized requirements downstream. Phase one (product planning) turns customer wants into product characteristics. Phase two (part deployment) turns those characteristics into part specifications. Phase three (process planning) turns part specs into process parameters. Phase four (production planning) turns process parameters into the control plans and work instructions the plant actually runs. The "hows" of one phase become the "whats" of the next, so a top customer priority stays traceable from a spoken want to a specific check on the line.

Most teams never run all four phases, and that is fine, the first house captures the bulk of the benefit. But the cascade is what makes QFD more than a prioritization exercise: it is a traceable chain from what the customer said to what the operator checks. That last matrix connects directly to statistical process control because the characteristics QFD flags as most important are exactly the ones worth putting on a control chart.

Where Does QFD Go Wrong?

QFD fails most often from scale, not from method. Teams build a house with forty customer wants and thirty characteristics, spend weeks scoring twelve hundred cells, and drown the signal in busywork. A useful house is small, a dozen wants, a dozen hows, focused on the decisions that are genuinely contested. The second failure is faking the voice of the customer: filling the importance ratings with the team's own opinions instead of real customer input, which turns the whole matrix into an elaborate way to confirm what engineering already believed. And the third is treating the matrix as the deliverable. The document is worthless if the priorities it reveals do not then change where design effort and design attention actually go.

By the Numbers

QFD is a documented, taught discipline with a clear lineage. ASQ describes the House of Quality as the primary planning tool used in QFD to translate customer requirements into engineering characteristics and prioritize them (ASQ, House of Quality), and the QFD Institute maintains the method Akao originated and continues to publish reference material and training on it (QFD Institute). The reason it endures is that it makes a normally invisible step, the translation from customer language to engineering targets, explicit and arguable, so the priorities can be checked instead of assumed. Where Harmony fits: QFD ends at the control plan on the floor, and that is where most of it is lost. Harmony keeps the characteristics that matter connected from spec to the line, so the priority the customer set does not quietly disappear between engineering and production, see the platform.