Key takeaways
- A Community Guidelines strike and a copyright strike share a word and almost nothing else. Different systems, different clocks, different appeals, and only one of them starts with a free warning.
- Your first violation is normally a warning with no penalty, and since August 2023 that warning can be cleared: complete the optional policy training and it expires 90 days later.
- The first strike costs you a week of uploading, live streams, Community posts and custom thumbnails — the thumbnail restriction is the one that quietly damages a video you already published.
- Thumbnails are judged under the same Community Guidelines as video, with a focal-point test rather than a rule about what can appear in frame. Repeat offences can remove custom-thumbnail privileges for 30 days.
- Deleting the offending video does not remove the strike — it removes your ability to appeal it. Appeal first, delete never.
- Three strikes in the same 90-day window ends the channel. Since October 2025 there is a narrow route back, and it is a new channel at zero subscribers, not a reinstatement.
Most creators meet the strike system in the worst possible way: an email at seven in the morning, a video gone, a countdown on the dashboard, and a week of upload restrictions starting immediately. At that point the useful reading is over. The decisions that mattered — whether to delete the video, whether to appeal, what the warning on your record actually meant — were all available beforehand and none of them are obvious under pressure.
This is the system explained while nothing is on fire. What a warning is and how to get rid of it, what each strike actually takes away, why the thumbnail is the asset most likely to trip a policy you have never read, and what the appeal windows are before they close. The copyright system is a separate machine with separate rules, covered in the piece on copyright claims versus copyright strikes; everything below is the Community Guidelines side.
Two systems, one word
The single most common mistake in creator advice is treating all strikes as one thing. They are not. A copyright strike comes from a rights holder submitting a legal removal request about your use of someone else's work. A Community Guidelines strike comes from YouTube deciding your content broke one of its own content policies. They run on separate counts, separate appeal routes, and separate remedies.
| Community Guidelines strike | Copyright strike | |
|---|---|---|
| Who triggers it | YouTube, via automated detection or human review | A rights holder, via a legal removal request |
| What it is about | What your content shows, says, or promises | Whose material you used |
| Is there a free warning | Yes, normally, for a first violation | No — the first one is a strike |
| How long it lasts | 90 days | 90 days |
| How to clear it early | A successful appeal | Retraction, counter notification, or expiry after Copyright School |
| Three of them | Channel termination | Channel termination |
They do share the ending. Three of either kind inside a 90-day window removes the channel. What they do not share is the path in, which is why the same creator can be careful about music licensing and completely exposed on thumbnails.
The warning, and the thing most creators do not know about it
Since February 2019, a first Community Guidelines violation normally produces a warning rather than a strike. The offending content comes down, and nothing else happens to the channel: no upload block, no feature restrictions. When YouTube announced that change it said 98% of creators had never received a strike at all, and that 94% of those who did receive a first strike never received a second — the case for the warning being that the system was catching people who make one mistake, not people who make a career of it.
The warning used to sit on your record permanently. In August 2023 that changed, and this is the part that gets missed. YouTube added an optional policy training: a short in-product course on the specific policy you broke, reported in the coverage at the time as around seven questions and roughly fifteen minutes. Complete it and your warning expires after 90 days. The 90 days run from when you finish the training, not from when the warning arrived, so the course is worth doing the day you get the notice rather than the week you remember it.
Two conditions sit around that. If you violate the same policy inside the 90-day window, the warning may not expire and the violation may become a strike instead. If you violate a different policy, you generally get another warning for that policy rather than an automatic escalation — warnings are tracked per policy, not as a single lifetime allowance. The practical translation: one careless thumbnail is survivable, and a habit is not.
The fifteen minutes nobody takes
A warning with no penalty feels like nothing happened, which is exactly why the training goes untaken. It is the cheapest insurance on the platform — a quarter of an hour that converts a permanent mark into one that disappears in 90 days, on the policy you have already proved you can trip. If there is a warning sitting on your channel right now from something you uploaded a year ago, it is still there, and it is still the thing standing between your next mistake and a strike.
What a strike actually costs
The strike itself lasts 90 days, as YouTube's documentation on strike basics sets out. The restrictions attached to it are shorter, and they escalate.
| Stage | What happens | How long |
|---|---|---|
| Warning | Content removed. No channel penalty. Policy training available. | Permanent unless training is completed, then 90 days |
| Strike 1 | Content removed. Posting and several creation features suspended. | 1 week of restrictions; strike lives 90 days |
| Strike 2 (within 90 days of the first) | Same restrictions, doubled. | 2 weeks of restrictions; strike lives 90 days |
| Strike 3 (within the same 90-day window) | Channel removed from YouTube. | Permanent, subject to appeal |
The restricted list during a strike is broader than "you cannot upload". It covers uploading videos, starting or scheduling live streams, creating premieres, posting to the Community tab, creating and editing playlists — and adding custom thumbnails.
That last item is the one that catches people who think a strike week is simply a week off. Custom thumbnails are a feature, and features are what get suspended. If a video is scheduled to publish during the restriction without a thumbnail already attached, it goes out with an auto-selected frame from the video — three algorithmically chosen stills, none of them designed, all of them competing in a feed against thumbnails that were. The same applies to the thumbnail you were planning to swap on an underperforming upload from last month. For a week, the packaging layer of your channel is frozen in whatever state you left it.
Worth being precise about what a strike does not do. It does not, by itself, remove an existing channel from the YouTube Partner Program. It does block a new application: if you have active Community Guidelines strikes you apply after they expire or after a successful appeal. And repeated violations are a different matter from a single strike — a pattern is exactly what the Partner Program's own policies are written to catch, even when each individual strike has expired.
The outcomes ladder, of which "strike" is only the top
Enforcement is not binary. A video that brushes a policy can land anywhere on a ladder, and most creators only ever notice the rungs that hurt.
- Nothing. The overwhelming majority of borderline content.
- Limited or no ads. The yellow icon — an advertiser-suitability decision rather than a Community Guidelines one, and a different system again, explained in the piece on the yellow monetisation icon.
- Age restriction. The video stays up but is unavailable to signed-out viewers and to under-18s, cannot be watched on most third-party sites, and — the detail that matters here — has its thumbnail blurred in search results. No strike is issued.
- Thumbnail removal. The image comes down, the video survives. A strike may or may not follow.
- Video removal with a warning. Content gone, channel untouched.
- Video removal with a strike. The week starts.
Rungs three and four are the interesting ones for anyone whose job is packaging, because both are outcomes where the video is judged acceptable and the image is not. An age-restricted video with a blurred thumbnail is still in search results, still nominally discoverable, and effectively unclickable. It is the closest thing YouTube has to deleting your click-through rate without deleting your video.
Why thumbnails are the exposed surface
YouTube's thumbnails policy is short and is not really a separate policy at all: content that violates the Community Guidelines is not allowed in thumbnails, and the same applies to banners, avatars, posts and every other image feature. The reason thumbnails generate disproportionate trouble is not that the rules are stricter. It is that the thumbnail is the one asset you deliberately designed to be noticed, made under time pressure, and often assembled from footage or images you did not shoot.
The policy works through a focal-point test rather than a list of forbidden objects. YouTube describes the considerations as whether breasts, buttocks or genitals are the focal point of the image; whether the subject is in a pose or clothing intended to sexually arouse the viewer; whether violent or gory imagery is the focal point; and whether written text is intended to be vulgar, or to shock or disgust. Focal point is doing the work in every one of those. A fitness channel filming a form breakdown is not the problem. A crop that centres the body part rather than the movement is a different image with the same footage behind it.
The four categories that catch ordinary channels
Almost all thumbnail enforcement against otherwise well-behaved channels falls into four buckets.
- Borrowed frames with content you did not vet. Reaction, commentary and news channels pulling a still from source footage. The clip may be fine inside a video with context; the still, standing alone in a feed with no context at all, is judged on its own.
- Shock text. Written text in the image is explicitly part of the test. The word that reads as edgy on a design canvas reads differently when it is the largest element on a 320-pixel-wide card next to children's content.
- Gore and injury as the subject. True crime, medical, combat sports, motorsport crashes. The video can be documentary in tone and the thumbnail still be the single most graphic frame in it, because that is how it got chosen.
- Regulated goods and dangerous acts. Firearms, drugs, vapes, gambling. The thumbnail is frequently what makes a video's subject unambiguous to an automated classifier that would otherwise have passed the whole upload.
There is a separate route in that has nothing to do with what the image shows. Misleading thumbnails are handled under the spam, deceptive practices and scams policies, which prohibit using titles, thumbnails or descriptions to trick viewers into believing content is something it is not. The examples YouTube gives are instructive: a thumbnail featuring a popular celebrity who has nothing to do with the video, or packaging that promises a genre the video never delivers. The severe version is what YouTube calls malicious clickbait, and the stated consequences there run to suspended monetisation or termination.
That line is not a ban on strong packaging. It is a deliverability requirement. Curiosity, tension, an implied question, a moment of visual drama — none of that is deceptive if the video contains the thing the image suggested. The tests worth applying are covered at length in the piece on the psychology of clickbait thumbnails, and the short version is that a viewer who feels lied to is a retention problem long before they are a policy problem.
The repeat-offence penalty specific to thumbnails
One consequence sits outside the general strike ladder. Where a thumbnail violates policy, YouTube removes the image and may issue a strike — and repeat offences can remove your custom thumbnail privileges for 30 days, or in serious cases terminate the account. Thirty days without custom thumbnails is not a warning shot. It is a month of uploads competing against designed images with auto-generated frames, on a platform where custom thumbnails have been a verified-channel feature for so long that the auto-frame is now a visible signal that something went wrong.
If you use AI to generate thumbnail images, the policy questions are slightly different and mostly about likeness rather than about disclosure; that is worked through in whether AI thumbnails are allowed on YouTube.
The hour after the email arrives
Enforcement notices reach you by email and as an alert in your channel settings. The full picture lives in YouTube Studio, on the dashboard's channel violations card, which is where the appeal button is too. Before you touch anything, there are three decisions in a specific order.
1. Do not delete the video
This is the single most expensive reflex in the whole system. Deleting the content does not remove the strike — the strike stays on the channel for its full 90 days — and once the video is gone you cannot appeal it. You have converted a decision you could still argue into one you cannot. The urge is understandable and it costs you the only remedy available.
2. Read the policy that was cited, not the policy you assume
The notice names a specific policy. Read that page before writing anything, because appeals are judged against the policy's own wording and examples, and the gap between the policy you were cited under and the one you think you broke is frequently where a winnable appeal is lost. A thumbnail removed under spam and deceptive practices is an argument about what you promised. The same thumbnail removed under a sensitive-content policy is an argument about what was in frame. They need different appeals.
3. Appeal inside the window, once
Under YouTube's appeal process, warnings and strikes can be appealed for six months from the date they were issued. Content removals — videos, posts, playlists, thumbnails and URLs — can be appealed for up to a year after the removal. You get the outcome by email, and there are three of them: the content is reinstated and the strike removed; the content is judged acceptable but not suitable for all audiences, so it comes back age-restricted; or the decision stands. Appeal a warning successfully and the next offence is treated as a warning again, which restores the buffer you thought you had lost.
Appeal when you have an actual argument — a misread image, a policy misapplied, context that the still does not carry. Appealing everything reflexively does not improve the odds on the one that matters, and each piece of content gets one appeal.
Termination, and the narrow door back
Three strikes in the same 90-day window removes the channel. So can a single severe violation, without any escalation at all — the ladder is for ordinary policy breaches, not for the categories YouTube treats as immediately disqualifying.
Until recently a termination was simply the end, including for the subscriber list and the back catalogue. On 8 October 2025 YouTube announced a pilot that changed this at the edges: some previously terminated creators can request a new channel. The terms are worth stating exactly, because the coverage tends to make it sound like reinstatement.
- You wait a year from the termination date before you can request one.
- The pilot excludes creators terminated for copyright infringement and those terminated under Creator Responsibility policies.
- Approval gives you a new channel starting at zero. Subscribers, watch history, monetisation status and the original channel identity do not come back. Old content can be re-uploaded if it complies with current guidelines, and the new channel applies to the Partner Program from scratch like any other.
The route that can actually restore a channel with its content and audience intact is still the appeal, available for a year after termination. The pilot is what exists after that has failed, and it is a second attempt rather than a restoration.
A policy you can run without thinking about policy
The point of understanding all this is to spend no further time on it. Four habits do almost all of the work.
Clear the warnings you already have. Open Studio, look at the channel violations card, and if there is a warning sitting there, take the training. Fifteen minutes against a 90-day expiry is the best trade available on the platform.
Run a five-question check on every thumbnail before it publishes. Is any body part the focal point rather than the person? Is the most graphic frame in the video also the image outside it? Does the text belong on a card sitting next to unrelated content? Is a face in this image someone who has not agreed to be in it? And the one that catches everything else: does the video contain the thing this image promises, early enough that a viewer finds it? That last question is a packaging question before it is a compliance one — the same standard the high-CTR rules arrive at from the opposite direction.
Keep thumbnails ahead of the schedule. The custom thumbnail suspension during a strike week only hurts if your images are made the hour before publishing. A channel with thumbnails attached to its scheduled uploads a few days ahead absorbs a strike week as an inconvenience rather than as a run of badly packaged videos. This is the same discipline that makes a consistent visual identity possible, and it pays off in both directions.
Check what the image looks like where it will be judged. Most policy-adjacent thumbnail mistakes look different at feed size than they do at full size on a design canvas — a crop that reads as a face at 1280 pixels reads as something else at 210. The thumbnail preview tool puts the image into the surfaces it will actually appear in, which is also where a reviewer, automated or human, first encounters it.
The honest summary
The strike system is more forgiving than its reputation and less forgiving than the moment it arrives suggests. There is a free warning, and since 2023 a way to erase it. The first strike costs a week, not a channel. Three inside 90 days is the end, and the route back is a new channel with none of your audience on it. Nearly every catastrophic outcome creators describe involves one of two errors: not knowing the warning could be cleared, or deleting the video and destroying the appeal.
What deserves more attention than it gets is that thumbnails sit inside all of this on equal terms with video. The image is judged by the same guidelines, it is judged standing alone without the context the video would have given it, it can be removed on its own, it can be blurred by an age restriction that leaves the video otherwise intact, and repeated problems can take custom thumbnails away for a month. For an asset that decides whether anything else on the channel gets watched, that is a lot of exposure to something most creators have never read a line of.
None of it argues for timid packaging. The channels that grow are still the ones with images that make a claim and videos that honour it. It argues for making thumbnails early enough to look at them twice, and for having more than one option so the strongest image is a choice rather than the only thing you had time to build. That production problem is what Thumblore exists to solve — several genuinely different concepts fast enough that the one you publish is the one that reads best at feed size, not the one the clock picked. If you want the enforcement picture completed, the copyright claim versus strike guide covers the other half of the system, and the yellow icon guide explains the monetisation decisions that look like enforcement but are not.