Visual-first openings are the largest group we found and the best performing on both measures

How this table was built, because it matters

Scrptly labels the opening of every reel it analyses. Those labels are written per video, so they are not a clean set of categories: across 2,518 Reels there are 2,405 distinct labels, and 2,360 of them appear exactly once. A ranking of raw labels would be 2,405 rows of one reel each and would mean nothing.

So the labels are grouped into nine families by keyword, first match wins. The grouping is ours, not a field in the data, and you should treat it as one reasonable cut rather than the only one. 472 Reels matched no family and are excluded from the table below, which is 19 percent of the sample.

The ranking

What holds up

Two findings survive the grouping. Visual-first openings are the largest family at 509 Reels, and they lead on both median views (183,376) and on the share beating three views per follower (22.5 percent). Number and list openings sit near the bottom on both, with 95,934 median views and the lowest clear rate in the table at 16.5 percent.

The gap between best and worst on median views is roughly three times. The gap on clear rate is about six percentage points. Neither is enormous, and with families this fuzzy you should not read small differences between adjacent rows as real.

What does not hold up

The tempting story here is that the analyser's hook score is inverted, rating the worst openings highest. It is not. Every family scores between 8.5 and 9.1, a range of six tenths of a point across nine groups covering two thousand Reels. The score barely discriminates between hook types at all.

That is still worth knowing, just as a different lesson: do not use a hook rating to choose between styles of opening. It is not measuring the thing that separates them.

What a visual-first hook actually is

Not a title card, and not a slow pan. It is a first frame that poses a question the viewer cannot answer by looking at it, and that the rest of the video answers. The setup is visible and the outcome is not.

What to do with this

1. Design the first frame before you write the first line. If the frame alone raises a question, the line can be shorter.
2. Be wary of opening on a number. It is the weakest family in the table on both measures.
3. Keep curiosity gaps. At 488 Reels they are nearly as common as visual-first and sit second on median views.
4. Judge your own opening by what it shows, not by how it reads in your notes app, and not by a score out of ten.

Figures on this page come from 2,518 Instagram Reels across 528 accounts, analysed by Scrptly. The source table holds 3,554 rows; it is deduplicated to 2,518 unique Reels so no reel is counted twice. Read 2026-10-08.