Key takeaways
- On 1 October 2026 YouTube said it is updating its Shorts recommendation systems to further prioritise original content and reduce the reach of channels that mainly aggregate or re-upload other creators' clips without significant changes.
- This is a distribution change, not a monetisation change. Reused-content rules already decided whether you could earn; this decides how often the feed shows you at all, which is the harder consequence to notice and the harder one to appeal.
- Creator Liaison Rene Ritchie was specific about what does not count: a reused clip is not made original by voice-over description of what is on screen, minor technical edits, or template-based bulk changes.
- What does count is value the source did not have — your commentary, your analysis, a different perspective, real editing, storytelling. The test is contribution, not effort.
- YouTube gave no start date, no re-evaluation timescale and no dashboard. There is no row in Analytics that says "reach reduced", so the diagnosis has to be inferred from impressions and feed traffic.
- Ritchie did say reach is re-evaluated if a channel shifts towards original content, which makes this recoverable in a way that a monetisation rejection is not.
For about five years, "is this original enough?" was a monetisation question on YouTube. You found out the answer once, at the Partner Programme review, and if you passed you stopped thinking about it. Channels built on other people's footage learned to clear that bar — a narration track, a caption overlay, a template — and then went back to publishing forty Shorts a week.
On 1 October 2026 the question moved. YouTube said it is updating the Shorts recommendation systems to put more weight on original content and to cut distribution for channels that mostly aggregate or re-upload other creators' clips without significant changes. Monetisation policy did not change. What changed is how often the feed puts you in front of someone, which is a quieter penalty and a much more expensive one. A demonetised channel knows it is demonetised. A down-ranked channel just thinks the algorithm is being strange this month.
This post works through what was actually said, what the stated bar for "original" is, which common formats sit on which side of it, how you would detect a reach reduction with no dashboard to tell you, and what to change first if the answer is that your channel is on the wrong side. It is also worth saying early: if you make your own videos and cut your own Shorts out of them, nothing here is aimed at you.
What YouTube actually said, and where
The announcement did not arrive as a policy rewrite. Reporting on the change describes it as a TeamYouTube community post, explained in more detail by Creator Liaison Rene Ritchie on the Creator Insider channel, and resting on the spam and deceptive-practices rules that already govern strikes, demonetisation and terminations. In other words: the standard did not move, the consequence did.
Three things in the announcement matter more than the headline. The first is the scope — recommendations. Coverage of the change is consistent that this is about how often YouTube surfaces a Short in the feed and elsewhere, while monetisation stays where it was, under the existing reused-content and inauthentic-content rules.
The second is the definition, and it is unusually blunt. Ritchie said that remixing someone else's clip requires adding value of your own, "not just VO descriptions of what's happening on screen, minor technical edits, or template-based bulk changes." Read that as a list of the three things the aggregation playbook actually consists of, named out loud.
The third is the recovery path. On what happens to a channel that has already been hit, Ritchie said: "If your channel was previously aggregating or re-uploading and you shift towards original content, the recommendation system will re-evaluate reach." That sentence is the reason this is worth acting on rather than mourning.
What YouTube did not give is any of the operational detail you would want. No date the new ranking starts. No timescale for how long a re-evaluation takes. No indicator anywhere in Studio that tells a channel it has been affected. Search Engine Journal's write-up makes the same point: creators were told the direction and left to infer the rest.
Two systems, and why the difference is the whole story
Almost every confused thread about this change comes from collapsing two separate machines into one.
| Monetisation eligibility | Recommendations | |
|---|---|---|
| What it decides | Whether the channel can earn at all | How often a Short is shown to someone |
| Governed by | Reused content and inauthentic content rules | Ranking systems, now weighted towards originality |
| How you find out | An explicit decision, with a reason | Nothing. Impressions are simply lower |
| Appeal | Reapply after 30 days | No appeal; re-evaluated if the content shifts |
| Changed on 1 October 2026? | No | Yes |
The practical consequence is that a channel can be fully monetised, in good standing, with no strikes and no warnings, and still be quietly carrying much less reach than it had in September. There is no message for that state. The only evidence is the shape of your own impressions.
It also means the usual advice — "appeal it", "contact support", "wait for the review" — does not apply. There is nothing to appeal. The input to the system is what you publish next.
The three things that explicitly do not make a clip original
Taking Ritchie's list seriously is the single most useful thing a clip-based channel can do this month, because each item maps to a production habit that felt like work.
Voice-over that describes what is on screen
This is the most common form, and the one creators defend hardest, because recording narration takes real time. But describing a clip adds no information a viewer did not already have from watching it. "And then he reverses straight into the bollard" over footage of a man reversing into a bollard is a transcript, not a contribution. The question the system is asking is not "did a human do something?" but "is there anything here the source clip did not contain?"
Minor technical edits
Cropping to vertical, bumping saturation, adding a progress bar, speeding the middle up by 10%, laying a trending audio track underneath. These change the artefact without changing the content. They are the edits you make when you have forty uploads to get out and nothing to say about any of them.
Template-based bulk changes
This one is aimed squarely at 2026 production stacks: the same intro, the same caption style, the same stock b-roll pattern, the same synthetic narration voice, applied across a batch of clips pulled from a feed. The template is the tell. If a pipeline can produce the next hundred uploads with no further decisions from a person, then nobody's judgement is in any individual upload, and that is precisely the thing being de-weighted.
Why "effort" is the wrong defence
Creators affected by this will point, reasonably, at how much work the pipeline was to build. That is true and irrelevant. The systems are not scoring hours spent; they are scoring whether the upload carries something the source did not. A fifteen-second cold open where you say the one thing nobody else noticed is worth more than nine hours of batch rendering, because only one of those is unavailable elsewhere.
What counts as value beyond the source
The positive half of the definition is where the useful instruction lives. As reported, incorporating clips from someone else's video is fine when you add original commentary, a different perspective, creative editing or storytelling — value beyond the source material. Four separable things, each of which can be produced deliberately.
Commentary is a claim about the clip. Not a description of it — a judgement, a correction, a piece of context the clip omits. The usable version is a sentence that could be argued with.
Perspective is the angle only you have. A mechanic watching a car crash clip, a nurse watching a medical drama scene, a pro player watching a bad play. Perspective does not need a script; it needs you to be the one holding the camera on your own reaction and knowing something.
Creative editing means the structure is yours: you cut three sources against each other to make a point neither makes alone, you build to something, you withhold and then reveal. Reordering someone else's video is not structure. Constructing an argument out of fragments is.
Storytelling is the one most clip channels never attempt, because it requires a beginning. A Short that opens with a setup, lands a turn and closes is doing something the raw clip cannot do, even if every frame in it came from somewhere else.
Where the line falls for eight real formats
Abstract definitions are easy to agree with and hard to apply, so here is the same test run against the formats people actually publish.
| Format | Where it sits | What moves it |
|---|---|---|
| Your own long-form, cut into Shorts | Original. Not in scope at all | Nothing — this is the use YouTube actively encourages |
| Your own TikToks or Reels, cross-posted | Original. It is your footage | Upload the clean export rather than a watermarked download |
| Reaction, filmed, with you on camera | Original, explicitly not the target | Keep your reaction the point; keep the source clip short |
| Commentary or analysis over someone's footage | Original, if the commentary makes claims | Replace description with judgement |
| Clips channel for a podcast you do not own | At risk — straight extraction, however good the taste | Permission from the show, plus framing, context or a through-line of your own |
| Compilations of other creators' clips | At risk, and the clearest case | A thesis the compilation proves, not just a run of clips |
| Templated faceless uploads with synthetic narration | At risk on two counts: reuse and bulk templating | Fewer uploads with research, scripting and a point of view |
| Re-uploading another channel's Short, unchanged | Squarely in scope, and a copyright exposure as well | Nothing. Stop |
Two rows deserve emphasis. The podcast-clips row surprises people because curation genuinely is a skill and good clippers demonstrably grow shows. It is still extraction unless you add something, which is why the healthy version of that business is a deal with the show rather than an unlicensed feed of it. And the faceless row is not an attack on faceless formats — plenty of faceless formats are entirely original without a face in them. What is at risk is not facelessness. It is templating.
The policies underneath, and what 2025 already told us
None of this appeared from nowhere. The monetisation side has carried an originality standard for years: reused content means material repurposed from elsewhere without significant original commentary, modifications, or educational or entertainment value — and that same clause is the escape hatch, because content built on other people's material does qualify when those things are present.
In July 2025 YouTube renamed the "repetitious content" guideline to "inauthentic content", which produced a week of panic about reaction channels being banned. TeamYouTube's response was that this was a minor update to a longstanding guideline, and Search Engine Journal's reporting at the time was explicit that reaction, clips and commentary channels were not the target. The target was content that follows a template with minimal variation, is easy to reproduce at scale, and has little evidence of an author in it.
Read 2025 and 2026 together and the direction is consistent. The standard has been the same for years; what keeps improving is detection, and what just changed is that failing the standard now costs reach rather than only revenue.
Why the feed, and why now
It helps to look at this from the position of whoever owns the Shorts feed rather than from the position of a creator. YouTube's last public figure, given by chief executive Neal Mohan in 2025, put Shorts at more than 200 billion views a day. At that scale the feed's problem is not supply of video. Supply is effectively infinite, and became more infinite the moment generative tools made it possible to assemble a hundred passable uploads in an afternoon.
When supply is infinite, the scarce input is anything that cannot be mass-produced, and the feed's only real asset is viewer trust that the next swipe is worth taking. Every re-upload that wins a slot does two kinds of damage: it spends an impression on something the viewer may have already seen, and it takes that impression from the person who made the thing. Shorts creators are paid out of a shared revenue pool — participating creators receive 45% of the revenue allocated from it — so distribution going to re-uploaders is not a neutral outcome for originals. It is a transfer.
That is the economic reading, and it is why this change is unlikely to be reversed. It is also why the right response is not to find the next thing that technically clears the bar. The bar is going to keep moving in the same direction.
How to tell whether you have been affected
With no dashboard and no notification, this has to be read off your own analytics — carefully, because 2026 has two traps that produce false positives.
The first trap is the counting change. From 24 August 2026 YouTube counts a public view the moment a video starts to play, with no minimum duration, across long-form, Shorts and live. It is not retroactive. Any comparison window that straddles that date is comparing two different measurements, and will read as a rise on the new side. If you are looking for a decline, that trap works in your favour — it masks drops — which is a reason to compare like-for-like periods after 24 August rather than year on year.
The second trap is the obvious one: Shorts performance is volatile at the best of times, and a single week proves nothing. Our post on Shorts that are not getting views works through the ordinary explanations, most of which are more likely than a ranking change.
With those out of the way, the signal to look for is specific. A recommendation change shows up as impressions, not as click-through rate and not as retention. Open Analytics, set a window entirely after 24 August, and look at Shorts feed impressions per upload across your last twenty or thirty Shorts. A reach reduction looks like a step down in impressions on uploads that are otherwise indistinguishable from the ones before them — same format, same quality, same cadence — with click-through and view duration roughly where they were. If impressions held and views fell, you have a packaging or opening problem instead, which is a different post: our guide to diagnosing a drop in views covers that funnel in order.
One more check worth running before you conclude anything: split your Shorts into the ones built from your own footage and the ones built from someone else's, and compare the two groups over the same window. If your original uploads are flat and your reused ones have stepped down, that is about as close to a confirmation as this change allows.
If you run a clip-based channel, what to change first
In rough order of return on the time it costs.
- Add a claim to the opening three seconds. Not a hook in the generic sense — an actual assertion about the clip that the clip does not make about itself. This is the cheapest conversion from description to commentary.
- Cut the source down. If the borrowed footage is 90% of the runtime, the ratio itself is the problem. Shorter source, more of you.
- Publish fewer, with research. A pipeline tuned for volume is the exact pattern being de-weighted. Ten considered uploads a month beat a hundred templated ones in a system that is now weighting originality.
- Break the template. Different structures, different lengths, different openings. If your last twenty Shorts are structurally identical, that is visible.
- Get permission where a relationship is possible. For podcast clips especially, an arrangement with the show changes both the rights position and what you are allowed to add.
- Make your own cover art. A frame lifted from someone else's video is the most literal possible signal of where the content came from.
Note what is not on that list: changing the thumbnail on existing uploads, deleting the back catalogue, or starting a fresh channel. Deleting rarely helps — the re-evaluation Ritchie described is driven by what you publish next, and there is a reasonable argument for leaving history where it is. If you are weighing a clean slate, be clear about what actually transfers to a new channel, which is less than most creators assume: not the subscribers, not the watch history, and not the goodwill.
Remixing inside YouTube's own tools
There is a legitimate, documented path for building on other people's material, and it is worth knowing where it sits relative to all of this. YouTube's Remix and Sampling features let you pull from an eligible video into a Short, and the result carries an attribution — a "Cut from [channel name]" link back to the source. Long-form videos are remixable by default and creators can turn it off per video in Studio, under the Shorts remixing setting. Shorts themselves cannot be opted out; the only way to withdraw one from sampling is to delete it, which also removes the Shorts built on it. Music-partner content is generally excluded.
Using those tools solves the rights question for the sampled material. It does not solve the originality question. An attributed remix with nothing added is still a reused clip as far as recommendations are concerned. Attribution tells the viewer where the footage came from; it does not tell the system what you contributed.
Copyright is a third system, and it has teeth
Worth separating, because people conflate it with both of the above. Reach reduction is not a strike. It leaves no mark on your channel's standing and does not affect uploads, live access or features. Copyright is a different machine entirely, with claims, strikes and actual terminations, and it is indifferent to whether your commentary was insightful. A compilation can be perfectly original by the recommendation standard and still collect three strikes. If you work with other people's footage at all, the difference between a Content ID claim and a copyright strike is the thing to be clear on before you publish, not after.
What this change does not mean
Four misreadings worth heading off before they set.
It is not a ban on using other people's clips. The announcement describes clips plus original contribution as fine. Commentary and reaction remain among the most durable formats on the platform.
It is not a strike or a penalty in the policy sense. There is no enforcement record, no notification and nothing to appeal, which is exactly why it is hard to detect.
It is not about AI as such. Generative tools are not disqualifying. Using them to produce volume with no authorial decisions in it is the pattern that gets caught, and that pattern predates the current tools by years.
It is not a long-form change. As announced, this is the Shorts recommendation systems. The monetisation standard on long-form has always been the same, and the broader direction of what YouTube announced at Made on YouTube 2026 suggests more of this logic rather than less.
Packaging still decides whether the reach you have converts
One consequence of a reach reduction is that every impression you do get becomes more valuable, which makes the packaging work matter more rather than less. Shorts have had genuine custom thumbnails since 25 July 2026 — Partner Programme channels, desktop upload, and notably no A/B testing on them, so the feedback loop is slower than on long-form. The mechanics are worth getting right, and our Shorts thumbnail guide covers where that cover is actually seen and how it differs from the first frame.
There is also an originality point hiding in the thumbnail. If your cover image is a frame from footage you did not shoot, it is doing the opposite of the job this update asks of you: it advertises the source rather than what you added. A cover built around your own face, your own type treatment or your own framing of the question is the version that signals a channel rather than a feed. That is the problem Thumblore exists to solve at the volume Shorts demands — generating covers in your channel's own style, from your own material, fast enough that making your own art stops being the reason you reach for someone else's frame.
Frequently asked questions
Has my reach already been cut?
Nobody outside YouTube can tell you, and YouTube has not said when the ranking change starts. The impressions comparison described above is the closest available answer.
Will a channel that was hit recover?
Ritchie said the recommendation system re-evaluates reach if a channel shifts towards original content. No timescale was given, so plan in months, not days.
Does this affect my monetisation?
Not directly — this is a recommendations change. Monetisation still runs on the reused-content and inauthentic-content rules, and Partner Programme earnings still run on engaged views rather than the public view count.
Are reaction videos safe?
Yes, as a format. A filmed reaction with you in it, adding your own response, is the paradigm case of adding value beyond the source. A "reaction" that is a clip with a muted presence in the corner is closer to the thing being de-weighted.
What about cross-posting my own TikToks?
That is your footage, so the originality question does not arise. Export clean from your editor rather than downloading a watermarked copy; a visible logo from another platform is a poor advertisement either way.
Does adding music or captions help?
Not for this. Both fall under minor technical edits. Captions are worth doing for accessibility and for sound-off viewing, but they do not change what the upload contributes.
Is curation really worth nothing?
Taste is real and good clippers do grow shows. But the system cannot see taste in a single upload, only what the upload contains. Make the curation visible — say why this clip, in your voice — and it becomes commentary, which does count.
The shape of the next two years
Strip the announcement down and it says something simple: the cheapest way to fill a feed is now the most penalised way, because the cost of filling a feed fell to nearly zero and the value of doing so fell with it. That logic does not stop at Shorts, and it does not stop in 2026. Every surface YouTube runs faces the same arithmetic.
Which makes the practical response narrower than it first looks. Not a workaround, not a format pivot, not a template that technically satisfies the wording. The durable position is to be the person who made the thing — or, when you are building on someone else's material, to be unmistakably present in what you publish. That is defensible against ranking changes in a way that no pipeline is, because it is the one input that cannot be reproduced at scale.
If your channel runs on volume, the next few weeks are a good time to read your own impressions honestly and decide which of your uploads would still be worth watching if somebody else had made them. For the ones that survive that test, the work is packaging them well enough that the reach you still have converts — which is where turning your own long-form into Shorts and a serious cover image earn their time back faster than another forty reposts ever will.