Trend advice tells you which opening formula to use. Across 2,078 Reels, the formula barely matters
Most trend advice is about opening formulas. Across 2,078 Reels the formula barely moves reach at all. What does, measured 23 September 2026.
What the numbers say
Almost every piece of trend advice is a formula for the first line. Start with a question. Use POV. Address the viewer directly. We sorted 2,078 Reels by how their opening line is grammatically built and measured the reach of each group. The formulas land within a whisker of each other. Reels opening with this or here reach a median of 0.52 times their follower count. Reels opening with a question word reach 0.42. Reels opening by addressing the viewer as you reach 0.34. That is the entire spread, from the best formula to the worst, across more than two thousand Reels. For comparison, on the same corpus, the measured quality of a hook separates 0.35 from 1.37, which is roughly four times. The shape of the sentence is close to noise. What is said inside it is not.
Copying the formula off a trend post is the cheapest thing to do and the least likely to change anything. If your Reels are flat, swapping every opener to a question will not fix it, because the data says that group performs slightly worse than average, not better.
Sample of 2,078, measured on 2026-09-23. Scrptly production database, queried 2026-09-23. 2,078 Reels deduplicated by shortcode, each with a view count, its account's follower count and a hook line beginning with a letter.
The measurement
Every Reel sorted by how its first line is built, not by what it is about. Views per follower is the Reel's views divided by its own account's follower count. Groups smaller than 60 Reels are not shown. The six bands below hold all 2,078: 155 plus 1,228 plus 300 plus 62 plus 225 plus 108. Read the median column first and notice how little it moves.
| How the line opens | Reels measured | Median views per follower | Top tenth reaches |
|---|---|---|---|
| With this, these or here | 155 | 0.52 | 9.08 |
| Some other way | 1,228 | 0.51 | 8.45 |
| With I, my or we | 300 | 0.49 | 7.75 |
| With a command or a format label such as POV or Day 1 | 62 | 0.44 | 9.76 |
| With a question word | 225 | 0.42 | 9.40 |
| By addressing the viewer as you | 108 | 0.34 | 8.14 |
One Reel, beat by beat
If the formula is not the thing, this is. One real Reel from the corpus, not a composite: 385,542 followers, seen 11,084,544 times, which is 28.75 times the following. It runs 59 seconds. Its opening line is a question, which the table above puts in the second weakest group, and it still went a long way. What carries it is the structure underneath, which sets a puzzle in the first six seconds and does not resolve it until the halfway mark.
| At | What happens |
|---|---|
| 0:00 to 0:06 | Set the puzzle: two versions side by side, one real and one not, and ask the viewer to pick |
| 0:06 to 0:10 | Hand off from the real person to the copy, on camera, so the puzzle is now visibly live |
| 0:10 to 0:26 | First act of the tutorial, shown as screen recording rather than described |
| 0:26 to 0:41 | Second and third acts, each introduced by a title card that acts as a reset for anyone drifting |
| 0:41 to 0:50 | Resolve the puzzle, and reveal the twist that the copy wrote its own script |
| 0:50 to 0:59 | Close on the real person reflecting, ending mid sentence, which is what drives the comments |
What to do next
- Stop auditing your openings by their shape. Whether a line starts with a question, a POV label or the word I is not what separated the Reels in the table above, and the spread between the best and worst formula is under two tenths of a view per follower.
- Audit them by whether they contain a reason to stay. The strongest performers in this corpus set something unresolved in the first few seconds and pay it off later, which is a property of the whole Reel rather than of its first five words.
- Treat a trending audio or format as a delivery vehicle, not a plan. It can carry a Reel that already has something to say. It cannot supply the something.
- One honest limit on all of this: these are Reels from accounts somebody chose to analyse, so they skew towards things that already worked. It is a fair guide to what strong Reels have in common and a poor guide to how often a given approach fails.
- Run your own handle through Scrptly if you want your openings scored rather than eyeballed. It returns the hook score and the beat-by-beat structure for the Reels that travelled. The first analysis does not need an account.