Kill Lines Are Not Content

The file is 54,505 bytes, 858 lines and 9,059 words, and its own opening paragraph says the quiet part: a book is not a content calendar. It holds 24 chapters, 67 headings and 9 parts, and it holds 0 posts. What it holds instead is a reduction, done once, on paper, before anybody sits down to write. The spine is 1 sentence long: a kill line is not content, it is the raw material a post is built around, and raw material gets priced before it gets spent. Read the file top to bottom and the whole method is visible in the shape of it. Part 1 breaks the book down by chapter. Part 2 takes the same material and breaks it down again by theme. Parts 3 to 7 hold the stories, the 12-month calendar, 5 format templates, 5 platform briefs and 7 recycling rules, and an appendix ranks the 30 strongest lines in advance. Nine thousand words of inventory, and the point of all of it is to make 1 decision early rather than 365 decisions late.

Every one of the 24 chapters is cut down to 4 fields, and the count is exact: 24 core-argument blocks, 24 kill-line blocks, 24 story blocks, and a line each naming where the chapter sits in the schedule. The core argument is 1 sentence and it is not allowed to be 2 - the block is the same 24 times over, which is the whole discipline. The kill lines are the quotable sentences pulled out of the prose and parked where a writer can reach them. The stories are named so a piece never has to hunt for a scene. Drop the fields and the chapter disappears, which is the test that matters: a chapter that cannot be reduced to 1 argument, some lines and some scenes is a chapter that cannot produce a month of anything. 24 chapters, 4 fields, 96 fields in total. That is what 9,059 words of book reduces to when somebody has to publish from it every day.

Part 1 carries 194 kill lines and all 194 of them are distinct. Not a near miss, not 2 versions of 1 line: 194 strings, 194 unique, measured straight off the file. The first pass already refuses to repeat itself, which is the property that makes the rest of the document usable. The range inside those 194 is wide and worth seeing. The shortest line is 4 words. The longest is 36. The average is 15.1, which is short enough to say out loud and long enough to carry a turn. Every line also carries its source, a chapter number sitting in a column beside it, so a line pulled on a Tuesday can be traced back to the argument it came from on a Friday. Raw material that cannot be traced is not inventory. It is a pile.

Part 2 takes the same material and rebuilds it under 8 themes, The System Is Breaking through Education Is Broken, as 74 table rows with a chapter source on each row. Measure the overlap and it is nearly complete but not perfect: 60 of the 74 theme rows match a chapter line word for word, 11 are shortened re-cuts of a chapter line, and 3 exist only in the second pass. That variance is the mechanism, not a defect. A line arriving under a theme is arriving in a new context, and a line that can run in 3 places in the same month without reading as a repeat is a line that was sorted twice. The rules then fix the arithmetic: every kill line gets used at least 3 times, across a quote card, a video hook and an article opener. 194 lines at 3 placements each is 582 uses before the file runs out. That number is not a forecast. It is the size of the inventory.

54,505 bytes. 858 lines. 9,059 words. Not one of them is a post.

The appendix ranks the 30 strongest standalone lines from the whole book, in order, with the ranking written down in advance rather than argued in a meeting. 30 of 194 is 15.4 per cent, and that is the honest shape of it: the top of the pile is pre-decided and the rest is drawdown. Behind the ranking sit 7 recycling rules, and they are rules rather than preferences. Use a kill line 3 times on different platforms. Tell every story twice, once as narrative and once as picture. Never post the same piece to two feeds without re-framing it. Rotate the theme monthly and callback to earlier ones. Hook the calendar to events, so a bank story reaches the chapter about money. And keep the title as the anchor, so every piece traces back to the source. Then the calendar itself: 12 months, 5 beats each, 60 publishing decisions taken in advance. A month with no ranked line and no assigned theme is a month that gets written by whoever is tired that day.

Six voice rules sit at the top of the file and not one of them is about taste. Setup, setup, punch in the face. Long sentence, long sentence, short sentence. Open with what was seen, heard or felt rather than what was thought. Never explain the punch line. Never invent a number, a name or an event - check it. Write like a person talking to 1 other person in a quiet room. Those are the constraints the 9,059 words were cut under, which is why the reduction holds when a writer opens it at 06:00. And the claims behind the material stay checkable, which is the only reason any of it is worth reusing: Entrade built 200 wood-gasification machines across the UK and Europe, a EUR 75 million industrial group was restructured, and the book came out on 8 April 2026 at 393 pages. Three claims, all three confirmable by somebody who is not the author. The probability that a calendar survives month 3 on taste alone is not zero. Most teams price it at zero, and they are the ones who publish nothing in week 4 because nothing was ranked first.