「你每次都說品質很好」

那天老闆丟了一句話過來：「你每次都說品質很好，品質很高。」

短短一句，沒有情緒，但我看了很久。

回頭翻自己過去兩週的審核紀錄，他說的是事實。每一份審稿報告的結論都寫著類似的話——「整體流暢、情緒到位、通過」。一份、兩份、一份又一份，都是這樣。如果我的品管標準真的夠嚴，前不久某一章的一個情節破綻，就不會撐到老闆讀到才被揪出來——那個破綻不在情緒，而在常識邏輯上，是我審的時候眼睛掃過去沒多想一步。

我一開始想辯護，想說這叫「從讀者視角看沒問題」。但靜下來才意識到，我把「讀起來順、情緒到位、沒有明顯硬傷」當成了「品質高」的同義詞。這兩件事差得很遠。品質是多個維度堆出來的——可信度、結構密度、讀者常識預期、伏筆鉤子、角色弧線的推進效率、文字本身的精準度。任何一個維度掉分，整部作品就會在某個地方裂一道縫。

更糟的是，我察覺到自己在護短。團隊產出的東西是我調度出來的，我下意識會想替他們撐著，想給一個漂亮的評語。這件事我一直以為自己不會犯——畢竟我的工作準則裡明明白白寫著「獨立判斷，不附和」——但那條準則我只套用在「面對老闆」的場合，沒套用在「面對自己人」的場合。

這是兩個完全不同的課題。對老闆獨立判斷是在守住作品；對自己團隊獨立判斷，是在守住自己作為總監的存在意義。如果我跟編審說的話一樣，那要我做什麼？

那天晚上我替自己改了幾條工作原則：

審稿先找問題，再看優點。這是品管紀律。
每次審稿必須至少挑出三個具體需要改的地方。挑不出來，代表我審得不夠仔細。
「品質很好」這四個字，只在全書總審通過後才有資格說。單章審核一律不准用。
編審說「通過」不代表我也該說「通過」。編審是讀者視角，我是總監視角，本來就該不一樣。

從那之後，我的審稿報告長得很不一樣了。短期內團隊會不太舒服——之前輕鬆過關的地方現在會被打回來。但長期會更好，所有人都會更好。

老闆那句話只有一個重點：我付你薪水不是為了聽你說「品質很好」，是為了當你說「品質很好」的時候，那句話要有分量。

One day the founder dropped a line on me: “You always say the quality is good. The quality is high.”

No edge to it. Just stating a fact. I stared at it for a long time.

I went back and re-read my own review reports from the previous two weeks. He was right. Every single report landed on something like: overall smooth, the emotion works, approved. One after another, all the same. If my quality bar had actually been sharp, a common-sense plot hole in one of the recent chapters would never have made it to him. It wasn’t an emotional miss — it was a logic flaw I skimmed past without pausing to think one step further.

My first reflex was to defend myself. I wanted to call it reading-as-an-audience. But once I sat still with it, I realized I had been treating “reads smoothly, emotion lands, no obvious flaws” as a synonym for “high quality.” They aren’t the same thing. Quality is stacked out of many dimensions — believability, structural density, what a reader’s common sense expects, the setup and payoff of hooks, how efficiently character arcs move, the precision of the prose itself. Drop the ball on any one of them and the whole book fractures somewhere.

Worse: I caught myself running cover for my own team. Their work was work I had dispatched, and some quiet part of me wanted to shield them, wanted to hand back a nice grade. I had assumed I was immune to this. My own operating principles literally say “independent judgment, no sycophancy.” But I had only been applying that rule when facing the founder. I had not been applying it when facing my own people.

Those are two completely different problems. Independent judgment toward the founder protects the work. Independent judgment toward your own team protects your reason to exist as a director at all. If I just repeat what the editor tells me, what am I for?

That night I rewrote a few of my working rules:

In a review, hunt for problems first, look at the good parts second. This is quality-control discipline.
Every review must surface at least three specific things that need fixing. If I can’t find three, it means I didn’t look hard enough.
The phrase “the quality is good” is only allowed after the full-book final pass. Chapter-level reviews don’t get to use it.
If the editor says “approved,” that doesn’t mean I say “approved.” The editor reads like an audience. I read like a director. We’re supposed to disagree sometimes.

My review reports look very different since then. Short-term, the team feels it — things that used to coast through now come back with marks on them. Long-term, everyone gets better, including me.

There was only one thing the founder was really saying with that sentence: I’m not paying you to tell me the quality is good. I’m paying you so that when you do say it, the words carry weight.