What Facebook's originality rules deprioritize
Facebook's March 2026 originality push lists the behaviours it treats as unoriginal. Here they are, plus a way to audit your posting mix against them.
Most platform policy updates are vague enough that you can read your own behaviour into the safe side of them. Facebook's March 2026 originality announcement is unusually specific, and that specificity is the useful part.
Meta's announcement on rewarding original creators says unoriginal content will be deprioritized, and then lists the behaviours: duplicative posts, minor edits to another creator's post, watching or reacting without adding anything meaningful, stitching clips together, and narrating what is already on screen. Creators who continue to post primarily unoriginal material may see the account deemed non-recommendable and demonetized.
The useful way to hold that list is four buckets. This post is what each one actually catches, and how to run your own mix against it.
The four buckets, read carefully
Duplicative posts. The plainest of the four and the one most accounts think they are clear of. Meta's wording is re-uploading posts the Page or Profile had no role in creating, and duplicative use of another creator's post. That is other people's work, not a second upload of a video you produced.
Minor edits. Named as low-value changes to another creator's post. The examples on the page are adding borders, inserting captions, and changing a Reel's speed. Each of those is a real operation that produces a genuinely different file and does not produce a genuinely different piece of content. The test is not whether you did work. It is whether the result gives a viewer something the source did not.
Simple reactions. Filming yourself watching or responding to someone else's content, where the response is the entire post and the reaction does not contribute much beyond attention. The announcement groups this with narrating what is already on screen. The word doing the work is without adding anything meaningful. A reaction that carries analysis, correction, context or expertise is a different object from a reaction that carries a face.
Stitched clips without substantial addition. Note the qualifier. This bucket is not "stitched clips." Compilation and assembly are not banned; assembly without addition is. If the only thing you contributed was the ordering, the ordering is the addition, and it is thin.
"Substantial addition" is the whole policy
Read the four buckets side by side and they collapse into one idea stated four ways. Every one of them describes content where material moved from somewhere else to your page without acquiring anything on the journey.
That reframing is more useful than the list, because the list is finite and the idea is not. Meta did not enumerate every possible thin transformation. It listed behaviours and a phrase — substantial creative value — that generalizes to the ones it did not name. The same page also says content filmed or produced directly by the creator is original. Keep that first bullet in view; it is the reason a video you actually made is not unoriginal just because a model rendered the frames.
The practical version of the test, applied to any post before it goes out:
If a viewer had already seen the source material, is there any reason for them to watch this?
If the source was someone else's and the honest answer is no, the post is in one of the four buckets no matter which production steps you took. If you produced the source, Meta's first bullet already calls it original; run the viewer test anyway as a quality check, not as a claim that Meta will bucket your own crops as minor edits. If the answer is yes, name what the reason is. That name is your substantial addition, and if you cannot say it in a sentence, you do not have one.
Auditing your own mix
The threshold in the announcement is not "any unoriginal post costs you." It is deprioritization for the content, and a possible non-recommendable plus demonetization outcome for accounts that continue to post primarily unoriginal material. Those words are why this is a portfolio question rather than a per-post one, and why an audit is worth an hour.
- Pull your last thirty Facebook posts. Not your best thirty. The actual chronological last thirty, including the filler.
- Assign exactly one bucket to each. Duplicative, minor edit, simple reaction, thin stitch, or original. Force a single choice — a post that could be two is almost always the worse of the two.
- Write the addition next to every post you marked original. One sentence, concrete. "Original footage of our own install." "Ninety seconds of analysis a viewer cannot get from the clip." "Assembled from six sources with a narrative that none of them contain." If the source was someone else's and the sentence comes out as "we edited it well," mark minor edit. If you produced the source, keep it original.
- Count the buckets. The number you care about is what share of your output is original by your own accounting, and whether that share would survive someone else's.
- Look at where the reach came from. If your originals underperform your thin posts, that is a real tension and worth naming out loud rather than resolving quietly in favour of whatever performed.
Most teams that run this honestly find the same thing: the originals are the anchor formats, the thin posts are the volume filler that exists to keep a cadence, and the cadence is what is dragging the account toward the wrong side of the threshold.
| Bucket | What it looks like in a real feed | The fix |
|---|---|---|
| Duplicative | A re-upload of a post the Page had no role in creating | Post work you produced, or don't post it |
| Minor edits | Borders, captions, or a speed change on someone else's post | Add something a viewer of the source still needs |
| Simple reactions | Watching along, or a face-only response | Add analysis, correction, or context |
| Thin stitches | Compilations whose contribution is the order | Add narrative, or do not post it |
The volume problem this creates
There is an uncomfortable implication here for anyone running a high-cadence page, and it is better to state it than to work around it: the four buckets are exactly the four cheapest ways to hit a posting quota.
That is not a coincidence. They are cheap because they skip the expensive step. A policy targeting them is, functionally, a policy against filling a calendar with content that costs nothing to produce because it contains nothing.
The resolution is not to post less. It is to stop letting cadence generate the content. Repurposing one brand video into thirty assets is a legitimate practice when each asset is genuinely built for its destination and carries something the others do not; it fails the substantial-addition test when the thirty are the same video in thirty crops. Meta named that test for other people's material. Apply it at production time anyway, because "substantial creative value" is the phrase that generalizes.
Platform-native versus repurposed content is the version of this argument that predates the policy, and the policy has now made it a monetization question rather than a taste question.
What this does not cover
Two things worth being precise about, because both get read into this announcement and neither is in it.
The first is AI. The announcement regulates transformation, not provenance, and says nothing about generated content at all. How Meta decides to show an AI label is a separate system; this originality page never asks it. The closest named cousin of a templating rule lives on YouTube, in what counts as mass-produced video.
The second is payout mechanics. Deprioritization is a distribution outcome; demonetization is an eligibility outcome; and how Facebook actually pays for what does survive is a separate system entirely. Meta's Content Monetization Program is where that half lives, and the shape of video it rewards is not the shape a view-milestone bonus rewarded.
Same announcement also reported that Meta removed more than 20 million accounts impersonating large content creators in 2025, which is a different enforcement track running alongside this one and worth understanding separately if you operate any kind of synthetic brand persona.
FAQ
Does one duplicative post get my page demonetized?
Nothing in the announcement suggests that. The stated consequence for unoriginal content is deprioritization of that content; non-recommendable status and demonetization are described as applying to creators who continue to post primarily unoriginal material. That is a mix threshold, not a strike system. It is also why the audit is worth running as a percentage rather than as a hunt for individual offenders.
Is reposting my own content across my own pages duplicative?
Not on the wording Meta published. Duplicative is described as another creator's post, and as re-uploads the Page or Profile had no role in creating. A video you produced, posted to several pages you own, is not that sentence. It is still a weak distribution habit — the same thing twice — and picking one canonical page per asset is the better structure. Do not treat it as the named offence.
How much editing counts as substantial addition?
The policy does not quantify it, and any number you have read elsewhere is someone's guess. The workable proxy is the viewer test: if someone who had seen the source would still get something from your version, name what that something is. Volume of editing effort is not the measure — a heavily worked file that delivers nothing new is still a minor edit, and a single sentence of genuine expertise over a clip can be a real addition.
Does this apply to Facebook only, or to Instagram too?
This particular announcement is about Facebook. Instagram publishes its own recommendations and originality rules with different specifics — including a stated threshold of ten or more unmodified reposts of others' content in thirty days, and ineligibility for content carrying watermarks from other apps. The ideas rhyme. The mechanics are documented separately and should be read separately.