故事 · 「要好」是骗不了车间的 Origin Story · "Make it better" doesn't survive the shop floor
质量屋一路从客户需求(WHAT)推到工程特性(HOW),看似已经很工程化了。可「降低噪音」「提高续航」 这些 HOW 仍然是方向,不是命令 —— 车间师傅一句「到底要做到多少」,就能把它问倒。 于是质量屋长出了最后一块、也最硬核的部件:地板(HOW MUCH)。它把每个方向钉死成一个数字: 噪音 ≤ 35dB、续航 ≥ 600km、重量 ≤ 1.2kg。有了数字,才有验收线,才有图纸,才有合格与否。 但定数字是门艺术 —— 定得太保守,做出来跟现状没差、被对手碾压;定得太激进,工程师怎么也够不着、项目延期。 好的目标值,是在客户期望、对手水平、技术可行这三股力量之间,找到那个既有竞争力、又真做得出来的点。 它一旦写进规格书,就既是对客户的承诺,也是套在工程师脖子上的约束。 The House of Quality marches from customer needs (WHAT) to engineering characteristics (HOW), and it already feels solidly technical. But phrases like "reduce noise" or "extend range" are still directions, not orders — one question from the shop floor ("OK, but how much?") and the whole thing falls apart. So the House grows one last, hardest piece: the floor (HOW MUCH). It nails every direction down to a number — noise ≤ 35 dB, range ≥ 600 km, weight ≤ 1.2 kg. Only a number gives you an acceptance line, a drawing dimension, a pass/fail verdict. But setting that number is an art. Too conservative and the product matches today's baseline — you get crushed by competitors. Too aggressive and engineering can never close the gap — the program slips. A good target lives at the one point where the three forces — customer expectation, competitor benchmark, and technical feasibility — leave room for something that is both competitive and buildable. Once it lands in the spec sheet, that number is at the same time a promise to the customer and a constraint around the engineer's neck.

1 目标值滑尺:夹在三条线之间 Target Sliders: Caught Between Three Lines

拖滑块试试Try the sliders

2 目标值要同时过三关 A Target Has to Clear Three Gates

客户期望线:达不到,客户根本不考虑你。这是入场券,是底线,不是目标。 Customer expectation line: miss it and you're not even on the buyer's shortlist. It's the entry ticket — the baseline, not the goal.
竞品水平线:达不到,就被对手压着打。目标值至少要够到、最好超过这条线。 Competitor benchmark: fall short of it and rivals dominate the comparison. Your target should at minimum match it — ideally beat it.
技术极限线:超过它,就是当前技术做不出来。目标定在它右边＝画饼,注定延期。 Technical limit: beyond it, today's technology simply can't deliver. Set the target past this line and you're drawing a pie in the sky — the slip is guaranteed.
黄金区＝竞品线右侧、技术极限线左侧的那段窄缝:既比对手强,又真做得出来。好目标值就该卡在这里。 The sweet spot = the narrow band right of the competitor line and left of the technical limit: better than the rivals, still actually buildable. That's where a good target value belongs.

3 现实里的工程目标值 Engineering Target Values in the Wild

可量化目标:「省油」落成「百公里油耗 ≤ 5.5L」,从此可测量、可验收、可争论。 Quantified spec: "fuel-efficient" becomes "≤ 5.5 L / 100 km" — and from that moment on it can be measured, accepted, and argued about.
HOW MUCH:质量屋地板专管「要做到多少」,把方向变成数字,把数字变成图纸。 HOW MUCH: the floor of the House of Quality owns the "how much" question — turning direction into a number, and that number into a drawing.
目标值权衡:手机防水做到 IP68 还是 IP67?多一档成本陡增,要在客户价值与代价间折中。 Target trade-off: phone water resistance — IP68 or IP67? One notch up multiplies the BOM cost. You have to balance perceived customer value against what it costs to deliver.
规格设定:目标值 + 容差 写进规格书,成为供应链、生产、检验共同遵守的硬约束。 Spec definition: a target value plus a tolerance lands in the spec sheet and becomes the hard constraint that supply chain, production, and inspection all live by.
一句话In One Line
目标值是质量屋里唯一一块没法含糊的部件。需求可以模糊地说「要好」,关系可以打个◎○△, 但地板上的数字不行 —— 它要写进规格书,要被供应商报价,要被质检卡尺量。 一个数字,同时是对客户的承诺、对工程师的命令、对生产线的标准。 定它的难,在于它被三股力量同时拉扯:客户期望往上推(要更好), 竞争压力往上推(要超对手),技术现实往下拽(做不到那么好)。 高手的本事,不是把目标定得最高,而是把它钉在那个又有竞争力、又真够得着的黄金点上 —— 定低了是平庸,定高了是空话,只有钉准了,才是能兑现的承诺。 A target value is the one part of the House of Quality that refuses to be fuzzy. Needs can stay vague ("make it better"), the relationship matrix can shrug with ◎○△, but the number on the floor cannot — it goes into the spec sheet, into supplier quotations, onto the inspector's caliper. A single number is at the same time a promise to the customer, an order to the engineer, and a standard for the production line. The difficulty is that three forces pull on it at once: customer expectation pushes it up (more please), competitive pressure pushes it up (beat the rivals), and technical reality pulls it down (we just can't do that yet). The mark of a great engineer isn't the highest possible target — it's nailing the target onto the one point that is both competitive and genuinely reachable. Aim too low and it's mediocre; aim too high and it's empty talk. Only a well-placed target becomes a promise you can actually keep.
常见误用Common Mistakes
目标只写方向,不写数字。「降低噪音」要落成「≤ 35dB」,无法量化就无法验收。 Stating a direction with no number. "Reduce noise" must land as "≤ 35 dB" — if you can't measure it, you can't sign off on it.
目标定得越高越好。超过技术极限就是画饼,会拖垮项目,要卡在可实现的黄金区。 "Higher is always better". Past the technical limit it's just wishful thinking, and it will sink the program — keep the target inside the feasible sweet spot.
只看客户期望,不查竞品和技术。三条线缺一不可,否则目标要么没竞争力,要么落不了地。 Considering customer expectation in isolation. All three lines have to be on the table — skip one and your target ends up either uncompetitive or impossible to build.

工程目标值