故事 · 为什么叫「房子」
Origin Story · Why It's Called a "House"
QFD 的第一张矩阵，因为长得像一栋带尖顶的房子，被亲切地称为「质量屋」。
它不是装饰，每一个部位都有职责：左墙列客户需求并标上重要度（谁更要紧）；
天花板列工程特性（我们能测量、能调的东西）；中间的房间是关系矩阵 ——
逐格回答「这个工程特性，对满足这条客户需求有多大帮助」；
三角屋顶记录工程特性彼此的相关性（提高这个会不会拖累那个）；
地板则把每列的关系强度乘以客户重要度再求和，自动算出每个工程特性的优先级。
一张图，把「要什么」和「怎么量、先做哪个」全对应了起来 —— 这就是 QFD 最有名的发明。
QFD's first matrix is affectionately called the "House of Quality" because it looks like a house with a pointed roof.
The shape isn't decorative — every part has a job: the left wall lists customer needs and their importance weights (who matters more);
the ceiling lists engineering characteristics (things we can measure and tune); the central room is the relationship matrix —
each cell answers "how much does this engineering characteristic help satisfy that customer need";
the triangular roof records correlations between engineering characteristics (does pushing one up drag another down);
the floor multiplies each column's relationship strengths by the customer importance weights and sums them up — automatically deriving the priority of every engineering characteristic.
One diagram, lining up "what the customer wants" with "how to measure it and which one to tackle first." This is QFD's most famous invention.
1 可交互质量屋：点格子，看地板重排 Interactive House of Quality: Click Cells, Watch the Floor Re-rank
点房间格子click a cell2 质量屋的五个部位 The Five Parts of the House
左墙 · 客户需求 + 重要度：客户要什么，每条要多重要(1~5)。WHAT。
Left wall · customer needs + importance: what customers want, and how much each one matters (1–5). The WHATs.
天花板 · 工程特性：可测量、可调控的技术指标。HOW。
Ceiling · engineering characteristics: measurable, controllable technical metrics. The HOWs.
房间 · 关系矩阵：每条需求与每个特性的关联强度 ◎○△。
Room · relationship matrix: the strength of the link between each need and each characteristic, scored ◎ ○ △.
屋顶 · 相关性：工程特性之间正/负相关，提示权衡冲突。
Roof · correlations: positive/negative correlations between engineering characteristics, flagging trade-offs.
地板 · 加权重要度：Σ（关系强度 × 客户重要度），自动得出工程优先级 —— 告诉你「钱该先砸哪个特性」。
Floor · weighted importance: Σ (relationship strength × customer importance), automatically yielding engineering priorities — telling you "which characteristic the budget should hit first."
3 现实里的质量屋 The House of Quality in Practice
质量屋：把客户语言和工程语言并排放进同一张表，逼着两边坐下来对齐。
House of Quality: puts customer language and engineering language side by side in one table, forcing both teams to sit down and align.
需求-特性映射：每个工程指标都能说清「服务于哪条客户需求」，没有无主的指标。
Need-to-characteristic mapping: every engineering metric can name "which customer need it serves" — no orphan metrics.
工程优先级：地板分数高的特性先攻关，把研发资源压在最影响客户的地方。
Engineering priorities: tackle the high-floor-score characteristics first, putting R&D resources where they hit the customer hardest.
竞品对标：右墙比对自家与对手在各需求上的表现，找出该追赶或拉开的差距。
Competitive benchmarking: the right wall benchmarks our performance against rivals on each need, spotting gaps to close or leads to extend.
一句话In One Line
质量屋最大的价值，不是那张漂亮的图，而是它强迫客户语言和工程语言坐到同一张桌上对齐。
市场说「要厚重的关门声」，工程问「那到底是声压级、频谱还是关门力」 ——
房间矩阵里的每一格，都是一次逼问与澄清。
地板的加权重要度则把主观争论变成可计算的优先级：
客户最在意的需求，自动把分数灌给和它强相关的工程特性，钱该先砸哪儿一目了然。
屋顶的相关性更点破了工程现实里的残酷权衡 ——
想更省油可能就得更轻，更轻又可能伤了关门的厚重感，鱼和熊掌得有人拍板取舍。
但要记住：地板的分数只和你填进去的重要度、关系强度一样可靠 —— 它是辅助判断的算盘，不是替你做决定的神谕。
The real value of the House of Quality isn't the pretty diagram — it's that the house forces customer language and engineering language to sit at the same table and align.
Marketing says "we want a solid door-close thunk," engineering shoots back "is that sound pressure level, frequency spectrum, or closing force?" —
every cell of the room matrix is one of those forced clarifications.
The floor's weighted importance turns subjective argument into computable priorities:
the needs the customer cares most about automatically pour points into the characteristics most strongly linked to them, making it obvious where the money should land first.
The roof's correlations expose the brutal trade-offs of engineering reality —
better fuel economy may demand a lighter body, but a lighter body can undermine the heavy thunk of the closing door. Somebody has to make the call.
But remember: the floor scores are only as trustworthy as the importance weights and relationship strengths you fed in — it's an abacus that supports judgement, not an oracle that replaces it.
常见误用Common Mistakes
把矩阵塞满几十行几十列。聚焦关键少数需求与特性，太大的屋子没人填得动也读不懂。
Cramming the matrix with dozens of rows and columns. Focus on the vital few needs and characteristics — an oversized house is unfillable and unreadable.
关系强度随手乱打，凑满格。◎○△ 要有依据，没关系就留空，地板分数才有意义。
Scribbling relationship strengths just to fill every cell. ◎ ○ △ must be evidence-backed; leave it blank when there's no real link — that's the only way the floor scores stay meaningful.
只算地板，忽视屋顶的负相关冲突。屋顶提示的此消彼长，正是最该团队拍板的权衡点。
Computing only the floor while ignoring the roof's negative correlations. The "one rises, the other falls" pairs the roof flags are precisely the trade-offs the team most needs to decide on.