I Checked Six Instagram Accounts' Recent Posts. Reel Ratio Tracks Archive Age, Not Reach.

I opened six public Instagram accounts on October 7, 2026 without logging in and recorded the twelve most recent items on each — seventy-two in total — classifying every item as either a Reel or a photo post.

The split came out at exactly half. 36 Reels, 36 photo posts. That number on its own tells you nothing. Sorted by account, it tells you a lot.

Six accounts, in order of archive size

106 thousand posts, sports league: 2 of 12 were Reels. Seventeen percent.
https://www.instagram.com/p/DeAXASAoE5H/

76 thousand posts, sports broadcaster: 4 of 12. Thirty-three percent.
https://www.instagram.com/reel/DdUj4ZQg8tQ/

32 thousand posts, news broadcaster: 9 of 12. Seventy-five percent.
https://www.instagram.com/reel/DeMGJ-Ek7Q4/

32 thousand posts, nature brand: 7 of 12. Fifty-eight percent.
https://www.instagram.com/reel/DdG4RIxIPyf/

5,890 posts, natural history brand: 10 of 12. Eighty-three percent.
https://www.instagram.com/reel/DbcP1_jjj7y/

4,952 posts, space agency: 4 of 12. Thirty-three percent.
https://www.instagram.com/p/DdHyaYAifb6/

The two accounts with the largest archives — 106 thousand and 76 thousand posts — are the two least likely to post video. The two with the smallest archives sit at the top of the list.

Why an old archive pulls toward stills

An account with a hundred thousand posts accumulated most of them before short video was the default format. Its grid, its highlights, its tagged-photo presence were all built around stills. Much of what an account like that still publishes is event coverage, promotional frames and graphics — formats that predate Reels and never needed to change.

A newer account carries no such obligation. Its first post could have been a Reel. There is no decade of imagery setting an expectation for what belongs in the grid.

This matters when you read recent output as a signal. On an old account, a low Reel ratio may be nothing more than inertia. On a new account, a high Reel ratio is closer to a deliberate choice, because nothing was pushing it there.

One more thing in the data

Caption length differs by format, in the direction you would not guess.

Reels in the sample carry a median caption of about 220 characters. Photo posts median around 186.

The video posts explain themselves more. You would expect the opposite — that a still image needs more words because it says less on its own. In this sample, the moving ones are the wordier ones.

Note that these captions are read from the image alt text Instagram serves to logged-out visitors, which carries the full caption on these accounts.

What I could not measure

Instagram does not expose like or view counts to a logged-out visitor. So this piece contains no engagement numbers at all, by design. I would rather report nothing than estimate a number someone might act on.

Method

Six public accounts, twelve most recent items each, read from the public grids on October 7, 2026 without logging in. Every item classified from its own URL — the platform marks Reels and photo posts differently.

Full notes with the original posts linked: https://reelvideosave.com/