Shadowban: real, imagined, and how to test it
The platforms say shadowbans don't exist. Creators swear they do. Both are sort of right: there is no secret switch, but there are several documented ways your reach gets cut without a removal notice. Here's what's real, what's a bad week, and a test that takes ten minutes.
Checked against the platforms' own help pages in August 2026.
Let's define the word first, because the argument is mostly about definitions.
"shadowban" in creator usage means my content is still up, nobody told me anything, and reach has collapsed. Under that definition, yes, it happens. Under the platforms' definition — a hidden, account-wide, punitive switch — no, and there's no evidence of one.
What is real
Every major platform now documents a state between "fine" and "removed": content that stays visible but is not recommended to people who don't follow you. The wording differs:
- Instagram: "not eligible to be recommended". Shown under Settings → Account → Account Status, together with a sample of the offending posts and a "Request a review" button. Affects Explore, Reels and feed recommendations; followers still see the content. Since 2025 and tightened in April 2026, accounts that mostly repost other people's content fall into this state for a rolling period and climb out by posting original work.
- TikTok: a video can be "not eligible for the For You feed". TikTok publishes its For You feed eligibility standards; the video's analytics show its status, and you can appeal from there. Account-level versions exist for spam-pattern behavior and repeated violations.
- YouTube: videos can be limited (age, "not suitable for most advertisers", flagged as made for kids) and channels can lose recommendation for reused or low-value content. Studio shows the state.
The honest translation of "shadowbanned" is one of three states — a content-level ineligibility (one video), an account-level one (your recent pattern), or a feature restriction (comments, hashtags, links). The platform will show you all three if you look in the right place; the full ladder is in how accounts get flagged.
What is imagined
Things that look like a shadowban and aren't:
- The first-test failure. A few hundred views and stop is the normal outcome for most videos (why). Three of those in a row is not a ban; it's three videos that didn't test well.
- The post-viral hangover. Reach after a big video drops back to baseline and feels like punishment (it isn't).
- Hashtag search. "My post doesn't show under the hashtag" — hashtag pages are curated and sampled; not appearing there proves nothing.
- Followers not seeing posts. Feeds are ranked; your followers see a fraction of what you post on every platform, always. This has been true for years.
- The friend who checked. One person not seeing one post in one feed is not a test.
A test that actually tests something
Ten minutes, no tools:
- Read the status pages. Instagram Account Status, TikTok video analytics and account settings, YouTube Studio. If there's a message, you're done — that's your answer and it's not hidden.
- Open your profile logged out, and from a second device. Are the posts there? Can a logged-out visitor play them? If not, it's privacy or removal, not reach.
- Compare followers vs. non-followers in analytics. Most platforms split reach by source. A reach that is almost entirely followers, for several posts in a row, is the signature of "not recommended to non-followers". A reach that is 90 % non-followers and just small is a test-batch story.
- Post something plainly original and plainly within the rules — a talking-head clip, your own footage, clean audio — at your normal time. If that also gets only followers, the account is in a reduced state. If it gets tested normally, the problem was the previous content.
What does not work as a test: searching your own username from your own phone, third-party "shadowban checker" sites (they query public endpoints and guess), or asking a group chat.
Getting out
If the status page names a reason, do the obvious thing: remove or fix what it points at, request a review if you believe it's wrong, and then post normal content at a normal cadence.
The platforms describe recovery in weeks, not hours, and they describe it as pattern-based. A month of clean, original posts is the mechanism — not a support ticket.
If nothing is named and the reach-by-source test says "followers only":
- Stop whatever was bursty — mass actions, rapid uploads, repeated comments, link spam.
- Stop reposting files that already exist on the platform (duplicate content). This is the most common unnamed cause for multi-page operators.
- Don't delete the archive. Mass deletion is itself an anomaly.
- Keep posting, at the same cadence, original material.
The operator's angle
If you run several pages, you will at some point have one in a reduced state while the others are fine. That's the argument for running them as genuinely different pages — own hook, own captions, own assets, own cut of the clip — rather than mirrors. A mirror that gets limited tells you nothing you can act on; a variant that gets limited while its siblings don't tells you exactly what the platform objected to. VidVertex renders per-brand variants for that reason (how variants work), and it logs which file went to which account so you can see the pattern (upload history).
Click to enlarge
Click to enlarge
Click to enlarge
"Shadowban" is a fine word for a bad feeling. It's a terrible word for a diagnosis. Go read the status page.