My writing room has a critic that isn't allowed to fix anything

Three weeks after the first episode of my show locked its script, I cut a line about my own background. It was wrong, and it had already been through every check I'd built.

That happened in the writing room, the third role on my one-person team. It takes an idea the research bench has checked and I've approved, and turns it into a script for The Operator Economy, my show about businesses one person could run. If you're just joining: every Monday this newsletter opens one box on that team's org chart and shows how it's set up, something it really made, what went wrong, and a template you can copy.

The job, in one sentence

Turn an approved idea into a script that's true, holds together as a business, and sounds like me talking.

The setup

On the org chart it's four boxes and a critic. One role works out how the business in the episode would make money. One keeps every claim inside what the research supports. One builds the story, and one writes the words I'll say. Then a review pass reads the finished draft looking for what's wrong.

The critic has one restriction I'd keep in any version of this: it can't fix anything. It writes findings, and that's all it does. If reviewers could edit, I'd end up with a script nobody actually wrote, and no record of why it changed.

Every finding gets a written answer: fixed, changed, rejected or kept on purpose, with the reason. Then I read the final script once, straight through, and answer one question. Does this sound like me talking? Nothing locks until that's a yes. Once it locks, every file gets a fingerprint, so any later change has to happen where I can see it.

The claims role usually earns its keep before a word is written. On that first episode, the most attractive number in the research, the one that would have opened the pitch, came from firms that sell the very service the episode was about. It was banned in every form, paraphrase included.

What it actually produced

Nothing new went through the room this past week. The show is down to one episode a month, so here are three rows from the review answers on that first episode, dated September 1.

• A scene opened with two sentences where the second didn't follow from the first, and the critic pointed out that an earlier fix had caused it. Accepted, and rewritten so the second sentence answers the first.
• The critic flagged "This is an M and A business" as reading oddly on the page. Rejected: the script is written to be heard, and spoken it's clear.
• The cold open promises "a business hiding inside it that almost nobody is running" before any evidence shows up. Kept on purpose, because the episode pays that promise off later, and moving it would take away the reason to keep watching.

There were ten findings, and each one got an answer like these. If you copy this, notice the rejected row most. Writing down why you said no is what stops the same objection from coming back every round.

What broke

The line I cut got my history wrong. I co-founded a company that exited. What I don't have is any experience getting someone else's business ready to sell, which is the job the episode is about. That part of the episode now opens with "I have no transaction experience," which is less flattering and correct.

Nothing in the room was set up to catch it. The claims role checks what the script says about the world against the research, and the critic reads for what's unclear or unsupported. A sentence about me looked like the one thing nobody needed to check. I caught it on September 23, while we were packaging the episode for release.

Two things made the fix clean. Because the script was fingerprinted, the cut shows up in the record instead of quietly disappearing. And the publishing runbook now says to settle every line about me with me before packaging starts.

A smaller one came from the review itself. Too many scenes in the draft opened with "So" or "Now", and fixing those openers broke the logic between two scenes. That was the first row above. Fixes make new problems, which is why the critic reads after the last round of fixes, not before.

Copy this

ROLE: Critic (the review pass)

JOB (one sentence):
Read the finished draft and list everything that's wrong with it.

INPUTS (the only places it may read):
- The draft
- The research and sources behind it
- My facts file

SCHEDULE:
Once per draft, after the last round of fixes

OUTPUT (same shape every time):
- Each finding: the exact line, what's wrong, why it matters
- No rewrites

LIMITS (it may never):
- Edit the draft itself
- Approve the draft
- Soften a finding because it's late

APPROVALS:
- Every finding gets a written answer before lock: fixed, changed, rejected or kept on purpose.
- I read the final version once, straight through. Nothing locks until I say it sounds like me.

If I were setting this up again, I'd add one input I didn't have: every line where the draft talks about me goes to me first, before the critic even starts.

Shipped this week

• The first emailed issue of this newsletter went out on Monday: the research bench.
• Dial Zero, the version of my assistant anyone can copy, now rehearses on a practice line, a pretend business that greets it, puts it on hold and sends it to voicemail, before it calls a real one.
• In the middle of a call it can also text its owner a question and wait for the answer, or ring the owner in and hand the call over.

Next issue, next Monday: the release gate, where nothing ships without a pass. If someone passed this along, you can get it every Monday at mpthaker.xyz.