we shoot 40 hooks. we ship 3.
the hook factory model, ratio and all. why we bench-test hooks before scripting the body, what our 8% ship-rate looks like across a year, and the spreadsheet that decides for us.
the hook is the product. the post is the packaging.
we shoot forty hooks for every three we ship. the body of the post is scripted only after the hook clears a three-step test. ship rate is roughly 8% across a year. we know this because we keep the spreadsheet that records every miss.
most reels are written hook-last. the writer maps the body, the script, the cta, then sits at the top of the page trying to back-fit a hook. it is the wrong order of operations and it produces, with great consistency, bodies that the hook can't earn.
we flipped it three years ago. we now do nothing else until we have a hook that has passed our internal test. most don't. the shoot day for any post starts with the hook bench, not the script.
— section onethe bench.
the hook bench is a recurring shoot block. forty hooks in three hours, six to ten seconds each. one writer, one director, one camera, two presenters who rotate. no scripts, no bodies, no plans for where the hook goes. just the first three seconds of forty different possible posts.
we don't try to be good. we try to be many. one in eight clears the test. that's the whole game.
by the end of the block we have forty raw hooks in a folder. we don't watch them back that day. we wait twenty-four hours. then we run the test.
— section twothe three-step test.
- 01the mute test. watch the hook with no sound. does it still ask a question. if no, dead.
- 02the cold-open test. show it to someone who doesn't know the brand. ask them: would you keep watching. binary answer. if no, dead.
- 03the body test. can you script a 30-second body that pays this hook off without lying. if no — meaning, the only payoff would be a bait — dead.
forty in. three out, on a good week. two out, on most weeks. some weeks one. we don't push the survivors. we don't lower the bar. the failures get archived in a folder we genuinely revisit, because some of them clear the test six months later when the context around them changes.
— section threethe spreadsheet.
we keep a spreadsheet. every hook gets a row. when it cleared, when it died, what killed it. "too quiet," "no question," "presenter blinked at second 1.2." over a year, patterns emerged. one presenter has a much higher clearance rate on opinion hooks. another has a much higher rate on demonstration hooks. we now cast the bench by hook type.
what the ratio looks like across the year
last calendar year we shot roughly 2,100 hooks. we shipped 168 posts. the ship rate was just under 8%. the average view count on the 168 was 4.6× what it had been the previous year, when we wrote hooks last. same team. same budget. same content pillars. the only thing we changed was the order of operations.
— section fourwhat this is not.
this is not a content factory argument. we are not arguing for volume. we ship roughly the same number of posts we did before — we just waste a lot less time on bodies that were going to die in distribution anyway. the bench is the cheap part. the body is the expensive part. you protect the expensive part by testing the cheap part first.
— closingthe question you should ask.
how many hooks did you write last week. not posts — hooks. if the answer is "one per post," your ratio is wrong, and your ship rate is hiding behind your assumption that everything you wrote was going to ship anyway. most of what you wrote should have died at the bench. some of it still will, on the platform, where the cost of failure is higher and the audience is a lot less polite.
this essay is part of the dispatch — one piece of writing on social, content, and the operations of running both, sent on the last friday of every month. no roundups, no listicles, no spam. subscribe here →