Instagram Reels Retention: A Guide for ChatGPT Users

If you’re checking Instagram Reels performance through ChatGPT specifically, you’re not alone, and the data backs that up: ChatGPT’s share of Instagram usage on Windsor.ai has grown substantially over the past two months, and Reels retention fields in particular get pulled through ChatGPT at a noticeably higher rate than ordinary post fields do. This guide is written with that in mind: how to check whether your Reels are actually holding attention, not just getting posted.

The four retention fields

Field	What it measures
`media_reel_skip_rate`	Share of viewers who skipped the Reel within the first 3 seconds
`media_reel_avg_watch_time`	Average watch time per view, in milliseconds
`media_reel_total_watch_time`	Total time the Reel was played, including replays
`media_reel_total_interactions`	Likes, saves, comments, and shares on the Reel, net of removals

Skip rate is the one worth checking first: a Reel with strong likes but a high 3-second skip rate is winning on the algorithm’s early signal despite a weak hook, which is a warning sign for its next few Reels, not just that one.

Catch a hook that isn’t working in the first three seconds

Pull skip rate and average watch time for my last 15 Reels. Sort by
skip rate, highest first, and show me the captions and post dates for
the 3 worst ones. What do they have in common in the first frame?

Compare watch time against your own history, not a benchmark

There’s no reliable external benchmark for Reel watch time, it depends too much on video length and niche. Your own account’s trend over time is the only honest comparison.

Show me average watch time and total watch time for all my Reels from
the last 90 days, grouped by month. Is watch time trending up or down,
and which Reel is the biggest outlier in either direction?

Find your best time to post Reels from your own data

Generic “best time to post” advice isn’t based on your account. Your own posting timestamps against your own view counts are.

Pull the timestamp and views for every Reel I've posted in the last 90
days. Group by day of week and hour posted, and tell me which posting
windows consistently get the most views.

FAQs

Why is this guide written for ChatGPT users specifically?

Usage data shows Reels retention fields get pulled through ChatGPT at a meaningfully higher rate than ordinary Instagram post fields, and ChatGPT’s overall share of Instagram usage on Windsor.ai has grown substantially in recent months. The prompts and fields here work the same in Claude and Copilot too.

What counts as a good skip rate?

There’s no reliable external benchmark, it varies too much by niche and video length. Compare a Reel’s skip rate against your own account’s recent average instead of an outside number.

Can I get a real "best time to post" answer from my own data?

Yes. Pull your own posting timestamps against your own view counts, that’s a real answer specific to your audience, unlike generic best-time-to-post advice based on aggregate industry data.

Do these fields work for regular video posts, not just Reels?

No. These four fields are Reels-specific. Regular image and video posts use the standard engagement fields instead.