Instagram Community Guidelines: What Gets a Post Removed vs Unrecommended

Instagram has two rulebooks, and the difference explains most "my reach collapsed" posts. The Community Guidelines decide whether content stays up, and breaking them produces a notification. The separate Recommendation Guidelines decide whether a post is eligible for Explore and Reels, and breaking those produces nothing at all.

Meta publishes both, and the recommendation set is deliberately broader than the removal set — content can be entirely allowed to stay and still not be shown to people who do not follow you (Meta Transparency Center).

That gap is the entire subject. Almost nobody's problem is removal.

If reach dropped and nothing was removed, work down this list.

"Tag a friend", "like if you agree", "comment YES", "double tap if". Meta has published a demotion for this for years and it is the single most common cause on this list, because it looks like ordinary caption writing.

It is also the easiest to fix: an actual question gets more comments than a demand for them.

The recommendation set covers things the Community Guidelines allow: content that is sexually suggestive without being prohibited, depictions of violence, regulated goods discussed rather than sold, health and financial claims, and misinformation flagged by third-party fact checkers. A post can be entirely within the rules and still sit in one of these bands.

"Guaranteed results", "cures", "lose ten pounds in a week", "financial freedom". This is the category most often tripped by people with no bad intent — it is standard advertising language and it is named in the published guidance.

A caption can be clean and the tag block not. Restricted and action-limited tags are covered separately in banned hashtags, and the tags people reach for when reach is already down are usually the worst offenders — see follow-for-follow hashtags.

Reach varies. A single post underperforming is not a signal, and the honest answer is that three or four consecutive posts are the smallest sample worth reacting to.

The caption is the part you can check before posting, and it is where the two most common causes live.

It reads a caption against the vocabulary behind each published category and says which category a flagged word belongs to. What it cannot do is judge the image, the audio or the context — moderation reads all three and a text checker reads text, so a clean result is a prompt rather than permission.

Three categories carry no word list at all: hate speech, harassment and misinformation. Those are decided by who a statement is about and what it claims, and a vocabulary for them would flag every anti-racism post on the platform while missing what is actually removed.

Run the tag block separately. It sorts a set into policy breach, sensitive category, long-reported restriction and merely saturated — four groups with different standing, and only the first is a fact.

There is no Instagram feature called a shadowban and the company has never described one.

What exists is the thing above: a post that stays up and is not recommended, with no notification and no appeal. That is a real mechanism, it is documented, and calling it by its name makes it fixable — you check the caption, the tags and the subject against a published list, rather than guessing at a secret one.

The same two-rulebook structure exists on TikTok, with a published list of content ineligible for the For You feed — see tiktok community guidelines. The general reach questions are in instagram growth.

Paste your caption into the community guidelines checker before you post, and run the tag block through the banned hashtag checker separately.