The Teardown Reads One Hundred

Most people watch a reel they like and try to copy the feeling of it. That is guessing with a stopwatch. The method that produces ground you can build on is duller, and it holds up: take 100 high performing reels, log every 1 of them against 9 fixed dimensions, then let the tally decide what you make next. The sample size is the method, not a detail of it. 10 reels is a mood. 20 reels is the floor at which a pattern can even be claimed, and below that you are describing your own taste back to yourself. 100 reels is where the repeats start to harden into a format, and where a flop stops being bad luck and starts being information. 1 spreadsheet row per reel. 1 column per dimension. No column invented in the moment, because the whole value is that the 5th reel and the 95th reel were read the same way, by the same columns, with the same questions. That is what turns a stack of clips into a benchmark instead of a mood board.

The 100 are not picked at random and they are not picked on vibes. Start from the 3 broadest phrases the audience actually types, and search those on 3 platforms, Instagram, TikTok and YouTube, so a phrase that only performs on 1 of them does not get mistaken for a law. Then balance the sample on purpose. 5 to 10 of the top niche creators, so the ceiling is represented. Rising creators under 100,000 followers who have 1 genuine breakout, because that is where a format is still being invented rather than defended. And the format that keeps appearing no matter who posts it, the shape that 4 unrelated accounts have already arrived at independently. 3 groups, 1 sheet. The ceiling tells you what is possible. The risers tell you what is working right now. The recurring format tells you what the platform itself is rewarding. Miss 1 of the 3 and the tally reads as a trend when it is only a cohort.

9 dimensions, and each 1 carries a fixed question. The hook. The structure. The pacing. The editing. The captions. The visuals. The length. The storytelling. The call to action. 9 columns, and the discipline is that nothing gets scored from memory after the fact. The hook column does the heaviest lifting and it holds 3 readings. The exact first frame words, their size and their placement. The first spoken line. And whether the register lands inside 1.5 seconds, meaning the viewer knows what kind of thing this is before deciding whether to stay. Visual, spoken and written all firing together is a different event from 1 of them carrying the hook alone, and the sheet should say which happened. Structure is logged as a beat arc: hook, context, first value hit, development, payoff, call to action, with the open loops marked where they open and where they close. Captions are logged at 36 to 60 point inside the safe zone, and words per frame are counted against the speech rather than trusted. Visuals are logged for lighting, depth, clutter and motion level, because a quiet frame and a loud frame do opposite jobs and the column should record which 1 was used.

Pacing is the column people fake, so it gets read against a hard number. 19 shots in the first 30 seconds, a change every 1.77 seconds, is the top creator signature. Log the average shot length, mark where the cuts cluster on the beat, and note the rhythm curve from the open through the middle to the close. Fast open, varied middle, is a shape. 1 speed for the whole thing is a tell. Editing is logged as decisions rather than effects. Hard cut, match cut, whip pan, zoom, speed ramp, snap zoom, text reveal, and the sound sting or the silence that lands with it. 1 line per reel is enough when the line is honest about what the cut did. Length is logged as an exact duration with its position in the niche distribution, so you can see whether the top of the niche clusters at 15 seconds or 45. And the call to action is logged as type and placement: follow, comment a word, save, share, link or none, early tease or mid or end, 1 ask or 3 stacked. A single ask and 3 stacked asks are different animals and the tally should never average them together.

100 reels across 9 dimensions is a sample with rules of its own.

Then the tally. Which hooks repeat in the top 20 per cent of the sample by views. Which lengths. Which pacing signature. Which call to action. That part is arithmetic, not interpretation. The section most people skip is the bottom 20 per cent, and skipping it is the mistake that costs the most. The flops read just as hard as the winners and they carry the better data, because a winner shows you 1 shape that worked once and a flop shows you a shape that does not work even when it is executed well. 2 pitfalls will hollow out the sheet. Fewer than 20 reels is noise pretending to be a pattern. And judging from the transcript alone misses the 2 things a muted viewer actually receives, the first frame and the captions, which between them carry half of the hook. If the sheet holds no frame data and no caption data, it has measured the radio, not the reel.

Every winner still misses the same thing, and that is the gap you are looking for. The niche has a proven format and no angle inside it that belongs to you. The proven format is the ground. The angle is the reason a stranger who has already seen the other 100 reels stops on yours. The output of the whole teardown is small and concrete: 10 to 20 one line ideas, each 1 built as a proven format times 1 difference. Not a format copied. A format used, with 1 thing the winners all left out. Here is the arithmetic this method sits on. Built companies across 12 countries. Restructured a industrial group. Deployed 210 energy systems across Africa and Asia. Took a company through a Frankfurt IPO. 16 years of field work from rural Nigeria to Fukushima. Entrade built 200 wood gasification machines. Signed a personal guarantee on a company that folded on Christmas Day 2008. Lost savings in the 2013 Cyprus bail-in. 393 pages, paperback and Kindle, published 8 April 2026. Every 1 of those is a body of work with an angle inside it that no competitor can copy, because it happened to 1 person and it took 16 years. The probability that a reel built on generic ground gets remembered is close to 0. The probability that the same craft, pointed at a proven format with 1 honest difference, gets remembered is close to 1. The difference is not the budget and it is not the camera. It is whether the format was chosen from a tally or from a feeling. 100 reels, 9 columns, 1 angle. That is the whole teardown.