故事 · 一张矩阵省下半年试错
Story · One matrix saves six months of trial-and-error
某团队同时看上四项技术:A 用现成方案就能做、还能立刻拉开和对手的差距;B 很炫但谁都做得出、做了也没人买账;
C 难到要啃三年、可一旦攻下就是护城河;D 既难又无关痛痒。
工程师本能地被 D 那种「有挑战的难题」吸引,差点把人力全压上去。
直到他们把四项画进一张 2×2 矩阵——横轴难度、纵轴优势——答案瞬间清晰:
先吃 A 这个快赢立住脚,用 A 的收益养 C 这个战略攻坚,B 和 D 直接划掉。
矩阵没给新信息,但它逼你把「酷」和「值」分开,这正是立项阶段最该想清楚的一件事。
A product team had four candidate technologies on the table at once. A could be built from off-the-shelf modules and would instantly differentiate them from competitors.
B was flashy but anyone could clone it, and customers didn't actually care. C would take three years of deep engineering — but once cracked, it was a moat.
D was both hard and irrelevant. Engineers, of course, gravitated toward D — the juicy technical challenge — and almost burned the whole headcount on it.
Then someone plotted all four on a 2×2: difficulty on one axis, advantage on the other. The answer landed in seconds.
Take A first — the quick win — to establish a base. Use A's revenue to fund C, the strategic bet. Kill B and D.
The matrix added no new information, but it forced the team to separate "cool" from "valuable" — and that's exactly the call kickoff is supposed to nail.
1 选型矩阵:拖动技术点看它落进哪个象限 Portfolio Matrix: drag a dot, watch its quadrant change
拖动彩点试试try dragging a dot2 该按什么顺序投:优先级一目了然 Funding order: priority at a glance
优先级 = 优势高 × 难度低。条越长越该先做。快赢排最前(投入小见效快),战略攻坚次之(值得长投),别碰类沉底。 Priority score = high advantage × low difficulty. The longer the bar, the sooner you start. Quick wins lead (small spend, fast payback), strategic bets follow (worth the long haul), skip picks sink to the bottom.
3 现实里的技术选型 Technology Selection in the Real World
快赢优先:先做高优势、低难度的项目立住脚、攒资源和信心,再去碰硬骨头——别一上来就赌最难的。
Quick wins first: ship the high-advantage, low-difficulty items first to bank resources and credibility — then tackle the hard stuff. Don't bet the farm on the toughest problem day one.
研发组合:用快赢的现金流,养几个长周期的战略攻坚,组合配置而不是把宝全押一处。
R&D portfolio: use quick-win cash flow to fund a couple of multi-year strategic bets. Run it like a portfolio — never bet everything on one ticker.
避开陷阱技术:又难又没差异化优势的技术最危险——它会以「有挑战」之名吞掉团队最好的人和最多的时间。
Avoid the trap quadrant: hard and non-differentiating is the most dangerous combo — it eats your best engineers and longest schedules under the cover of "interesting challenge".
别为鸡肋花力气:容易但谁都能做、做了也不带来优势的,做不做都行,别占用关键资源。
Skip the filler: easy but undifferentiating — anyone can build it, customers won't reward you for it. Do it only if it's free; never burn critical resources on it.
一句话In One Line
工程师的本能是被「难」吸引——越难的题越想征服。但立项不是解谜大赛,
难度本身从不创造价值,只有「优势」才创造价值。
这张矩阵的真正作用,是把「我想做」和「值得做」强行分开:纵轴先问做成了对客户、对竞争有什么用,
横轴再问要付出多大代价。两者一交叉,资源该流向哪里就不再靠拍脑袋。
记住那条对角线:同样的优势,越省力越优先。
Engineers are wired to chase hard problems — the harder, the more tempting. But kickoff is not a puzzle competition.
Difficulty alone never creates value; only advantage does.
The real job of this matrix is to force a clean split between "I want to build it" and "it's worth building". The y-axis asks first: what does it do for the customer and the competitive position?
The x-axis asks next: what does it cost us to ship? Cross the two and resource allocation stops being a gut call.
Remember the diagonal: for the same advantage, lower effort always wins.
常见误用Common Mistakes
被「技术挑战」诱惑,专挑最难的做。难度不产生价值,先按优势排序,再在同等优势里选省力的。
Picking the hardest problem just because it's hard. Difficulty doesn't pay the bills. Rank by advantage first; within the same tier, take the cheapest path.
把「我们能做」当成「值得做」。能做 ≠ 有差异化优势;左下角的鸡肋做了也白做。
Mistaking "we can build it" for "we should build it". Capability ≠ differentiation. The bottom-left filler ships and changes nothing.
阈值随便定,象限失去意义。优势/难度的分界要结合市场和团队真实能力校准,不是固定 50。
Picking thresholds arbitrarily, killing the matrix's signal. Calibrate the cutoffs against real market data and honest team capability — 50 is a placeholder, not a rule.