Key takeaways
- On newly uploaded long-form videos you no longer choose ad formats. YouTube removed the separate pre-roll, post-roll, skippable and non-skippable tick-boxes for new uploads in late 2023, leaving mid-roll placement as the one ad lever that is still yours. Videos uploaded before the change kept their old settings.
- Since 12 May 2025 YouTube places automatic slots at natural break points — pauses and transitions rather than mid-sentence — and it backfilled automatic slots into older videos that had only manual breaks. There is an opt-out in Studio, and Studio now tells you when one of your own manual slots looks interruptive.
- YouTube's own experiment found channels running automatic slots alongside manual ones earned on average over 5% more ad revenue than channels placing breaks by hand alone. That is the only first-party number in this subject worth planning around.
- A break only earns from viewers who are still watching when it arrives, so the retention curve, not the timeline, decides what each slot is worth. Two breaks in the first half of a video usually beat four spread across the whole of it.
- Since March 2026 a viewer on a television may be served a 30-second non-skippable ad that mobile viewers never see. The same badly placed break now costs considerably more attention than it did two years ago.
- Ad placement is a revenue-per-view lever worth low single-digit percentages. Packaging is the lever that decides how many views exist at all. Fix the breaks in half an hour, then go back to the thumbnail.
Ad settings are the part of Studio most creators configured once, in a hurry, during the week they got monetised, and have not looked at since. That was a defensible choice when the panel was full of switches nobody could evaluate. It is a worse choice now, because almost all of those switches have gone and the one that remains has quietly become more consequential.
Two changes did it. YouTube took format selection away from creators on new uploads, so whether a viewer gets a six-second bumper or something considerably longer is decided by Google's systems rather than by you. Then, in March 2026, YouTube began serving 30-second non-skippable ads on connected televisions worldwide. Put those together and the picture is clear: you cannot choose what the ad is any more, but you still choose the moment the viewer meets it, and the stakes attached to that moment have risen.
This is the working guide to that one remaining control. What the settings actually are and where they live, how automatic placement behaves since the 2025 rewrite, the only credible research on where ads survive, how to read your own retention graph to decide where breaks belong, and an honest sizing of how much any of it is worth next to the packaging work that produces the views in the first place.
What you actually control in 2026, and where it lives
Start by separating the levers from the illusions. The panel still looks like a set of choices, but most of what it shows is reporting rather than control.
| Lever | Still yours? | What it does |
|---|---|---|
| Monetisation on or off, per video | Yes | The blunt switch. No ads at all on that upload |
| Manual mid-roll slots | Yes | You pick timestamps in the break editor |
| Automatic mid-roll slots | Yes, a tick-box | YouTube adds slots at points it judges natural, on top of yours |
| Bulk mid-roll settings across the catalogue | Yes | Apply automatic breaks to videos without them, or to everything |
| Pre-roll, post-roll, skippable, non-skippable | No, on new uploads | Decided by YouTube since the late-2023 change |
| Which ad length a given viewer sees | No | Selected by device and context, including the 30-second TV format |
| Advertiser suitability | Indirectly | Self-certification and the guidelines decide whether you get full or limited ads |
The last row is the one creators most often mistake for an ad setting. If a video is running with the yellow icon, no amount of break placement will fix its revenue, because the problem is the size of the advertiser pool rather than the number of slots. That is a different diagnosis with a different fix, covered in the piece on limited ads and the yellow icon.
In Studio on desktop, open Content, click the video, then Monetisation. YouTube's help documentation for managing mid-roll ad breaks describes the mid-roll section there: you place breaks on the timeline yourself, and a tick-box controls whether YouTube adds automatic slots as well. By default automatic slots sit on top of your manual ones, so unticking the box is the only way to get a strictly manual video.
For the back catalogue, the same documentation points at the bulk path: select videos in Content, click Edit, then Ad settings, and choose whether mid-rolls apply to videos that currently have no breaks or to everything selected. This is the fastest way to correct a channel-wide mistake, and the fastest way to make one, so read which of the two options is selected before confirming.
Ads can also be switched off for an individual video. That remains useful for a sponsored upload where a pre-roll from a competing brand would be awkward, for a community message you do not want monetised, and for anything you are about to put in front of a client.
How the controls narrowed, in four steps
The current setup is the result of four separate decisions, and knowing which one produced which behaviour saves a lot of confused forum reading.
| When | Change | What it left creators with |
|---|---|---|
| July 2020 | Mid-roll eligibility lowered from ten minutes to eight | The eight-minute line that still shapes long-form run times |
| Late 2023 | Format tick-boxes removed on newly uploaded long-form videos | Mid-roll placement only; older uploads kept their settings |
| 12 May 2025 | Automatic slots moved to natural break points; auto slots added to older videos with manual breaks | An opt-out, a hybrid mode, and Studio feedback on interruptive slots |
| March 2026 | 30-second non-skippable ads launched on connected TVs | No new control, a higher cost for a break in the wrong place |
Reporting on the 2023 change noted that few creators had been using the format tick-boxes anyway, which is probably true and also beside the point. The direction of travel is what matters: every revision since 2020 has moved a decision from the creator to the system, with the single exception of where the breaks sit. That exception is where your attention belongs.
A break is worth whatever your retention curve leaves standing
Here is the mechanism that most break-placement advice skips. An ad slot does not earn money because it exists. It earns when a viewer is present at that second of the video and the impression is served. So the value of a slot is a function of two things: the price of the impression, which you do not control, and the share of your audience still watching at that timestamp, which you can read off a chart you already have.
Open the retention curve for a recent upload. If you are not fluent in reading one, the retention graph walkthrough covers the shapes and what they mean. Then do the arithmetic that the Studio interface never shows you. Imagine a twenty-minute video where, as is entirely ordinary for long-form, roughly half the audience is still there at four minutes, a third at ten, and a fifth at seventeen. Those proportions are an illustration rather than a measurement — yours will differ — but the shape holds for almost every channel, and the consequence is uncomfortable for anyone who spaces breaks evenly.
Under that curve, a break at four minutes reaches roughly two and a half times the audience of a break at seventeen. Two breaks in the first half therefore capture most of the impressions that four spread evenly across the video would, at half the interruption, because each later slot adds less than the one before it while costing the same patience. Evenly spaced breaks feel fair and are the arrangement that buys the fewest additional impressions per interruption.
Two caveats keep this honest. Impression prices are not constant across the video, and the audience still present at seventeen minutes is more committed than the audience at four, which may make them worth more per impression to some advertisers. And a slot that sits in the first minute competes with the pre-roll the viewer has just sat through. The point is not that late breaks are worthless; it is that a timeline is the wrong map. The retention curve is the right one.
The question the reports cannot answer
Studio will tell you a video's revenue, its RPM and its playback-based CPM. It will not tell you what any individual break earned. You cannot therefore A/B test slot positions the way you can test a thumbnail, and nobody who claims a precise per-slot figure for your channel is reading it off a report. Treat placement as a decision made from mechanism and retention shape, not as an experiment with a readable result.
What the research actually says about where ads survive
There is remarkably little public data on ad position, and what exists is old. The most useful study remains Shankar Krishnan and Ramesh Sitaraman's 2013 measurement study of video ads, which analysed anonymised traces from Akamai's video delivery network covering around 65 million viewers, 362 million videos and 257 million ads across 33 video providers. Three findings from it bear directly on placement.
An ad was 18.1% more likely to be completed as a mid-roll than as a pre-roll, and a pre-roll was 14.3% more likely to be completed than a post-roll. An ad in long-form content such as films and television episodes was 4.2% more likely to complete than the same ad in short clips. The ordering is intuitive once you think about sunk cost: a viewer who has already invested ten minutes is far less likely to walk away from thirty seconds than a viewer who has invested nothing.
The limits matter as much as the findings. The dataset predates YouTube's current ad stack, covers a delivery network rather than YouTube specifically, and measures ad completion rather than the thing a creator cares about, which is total watch time across the session. Read it as evidence about human patience, which does not change much, rather than as a model of YouTube's auction. What it supports is a modest claim: the mid-roll is the position where an interruption is best tolerated, and the tolerance grows with how much the viewer has already invested.
Why YouTube started moving your slots for you
On 12 May 2025 YouTube changed how automatic mid-rolls are positioned, aiming them at natural break points — pauses and transitions — and away from the middle of sentences and action, as reported at the time. Three parts of that release are easy to miss and still relevant.
First, the backfill. Videos uploaded before 24 February 2025 that carried only manually placed mid-rolls had additional automatic slots inserted at natural points. If your ad load on older uploads looks heavier than you remember setting, that is why, and it was accompanied by an opt-out in Studio for creators who want their manual placement left alone.
Second, the number. YouTube said an experiment run the previous year found channels using automatic slots in addition to manual ones earned on average over 5% more ad revenue than channels relying on manual placement alone. It is a modest, specific, first-party figure, and it is close to the only one available in this area. It is also an average across channels with wildly different content, so it is a prior rather than a prediction.
Third, the feedback. Studio gained the ability to tell you when one of your manual slots is judged interruptive. That is the most useful part of the release and the least discussed, because it turns a guess into a check. If you place breaks by hand, the flag is worth reading before publish rather than after.
It is worth being clear about whose interest this serves. YouTube's incentive is a viewing experience good enough to keep sessions long, which overlaps heavily with yours but is not identical to it: the platform optimises across every video a viewer watches, while you care about one. In practice the overlap is large enough that its natural-break detection is a better default than most creators' instincts, which is an unglamorous thing to admit about an automated system and appears to be true.
Automatic, manual, or both
The right answer depends on how structured your edit is. A video with clean topic boundaries gives a human placer an advantage that automation cannot match; a loosely edited one does not.
| Format | Recommended mode | Reasoning |
|---|---|---|
| Scripted essay or documentary | Manual, then check the flag | You know where the act breaks are; an ad over a reveal is expensive |
| Tutorial with distinct steps | Manual at step boundaries | Steps are natural seams and viewers resume easily |
| Podcast or long interview | Automatic, or hybrid | Few hard boundaries over ninety minutes; detection beats guessing |
| Vlog | Hybrid | Place one or two by hand at scene changes, let automation fill the rest |
| Livestream VOD | Automatic | Nobody is hand-placing breaks across a four-hour replay |
| Eight to ten minutes total | One break, manual | Room for exactly one that is not in the hook or the payoff |
| Back catalogue | Automatic in bulk | The revenue per video is too low to justify hand-placing hundreds |
Hybrid is the default worth starting from, because it is the arrangement YouTube's own experiment measured favourably and because it fails gracefully: your deliberate break at the act change survives, and the slots you did not think about land somewhere defensible rather than nowhere.
Five places a break does real damage
Placement advice is mostly a list of exclusions. Rule these out and the remaining candidates are usually fine.
- Anywhere in the opening. The first thirty seconds are doing the work of keeping the viewer at all, an interruption there competes with the hook itself, and a viewer who leaves during an ad in the first minute never reaches the rest of the video. The opening is the least monetisable stretch of any upload precisely because it is the most valuable.
- Immediately before the payoff. If the video has promised something, the moment before the delivery is the point of maximum tension and the worst possible interruption. It also reads as deliberate hostage-taking, which is the one placement viewers comment on.
- Inside a sponsor read. You have sold that attention to someone. A pre-roll from a competitor appearing inside their segment is a conversation you do not want with a brand partner.
- Mid-sentence or mid-action. This is exactly the pattern the 2025 change was designed to reduce, and if automation is now avoiding it, a manual slot that does it stands out worse than before.
- Over the end screen and call to action. The final stretch is where the next-video click and the subscribe happen. Trading that for one low-reach impression is a bad trade in both directions, and it is a common accident on videos where the editor added breaks by eye.
Designing seams while you edit
The creators who place breaks well are not better at the Studio timeline; they build the seams during the edit. Three habits do most of it.
Write the breaks into the structure. If a script has four sections, it has three natural seams, and a one-second beat of silence or a transition at each gives both automation and a human somewhere obvious to land. This costs nothing and makes the video better paced regardless of ads.
Use chapter boundaries as candidates. Chapters already mark the points you considered structurally meaningful, which is most of the work of identifying a natural break — see the piece on chapters and key moments for how they are generated and what they affect. A break on a chapter edge is also the version a returning viewer tolerates best, because the resume point is a thing with a name.
Place from the retention curve of your previous upload rather than from the current timeline. If your videos reliably shed a third of the audience in the first two minutes and then hold steady, your break candidates sit after that cliff, not before it. Patterns in retention shape are stable enough across a channel's uploads to plan with.
The eight-minute line, and the padding trap
Mid-rolls require a video of at least eight minutes, a threshold lowered from ten in 2020 and unchanged since. It is the single most distorting number in YouTube monetisation, because it tempts creators to stretch a six-minute idea past the line.
That trade is almost always bad, and the mechanism is the same retention arithmetic as above. Padding adds minutes that shed audience, which reduces average view duration, which reduces the impressions the recommendation system is willing to hand you in the first place. You gain the right to place a break and lose some of the reach that makes breaks worth anything. The revenue side of that trade is worked through in what YouTube actually pays per thousand views. The short form: clear eight minutes when the material genuinely fills it, and accept pre-roll and post-roll only when it does not.
The living-room change raises the cost of a bad break
From early March 2026, YouTube began rolling out 30-second non-skippable ads on connected televisions globally, a format that appears on TV apps rather than on phones and laptops. Coverage at the time, including 9to5Google's report, described Google's systems selecting between six-second bumpers, 15-second spots and the 30-second connected-TV format depending on the viewer's device and context. Creators do not choose between them.
For break placement this changes the arithmetic of a mistake. A skippable ad on a phone costs an impatient viewer a few seconds before the skip button does its job. For a television viewer in 2026 the same break may cost thirty seconds they cannot skip at all. Televisions are also where a growing share of long-form watch time now happens, which is why the television-first considerations keep surfacing in packaging advice as well. The practical upshot is simple: the margin for error on a break that interrupts something has narrowed, and the case for fewer, better-placed breaks is stronger than it was.
How much is any of this worth?
Be realistic about the size of the prize, because the internet is not. The one first-party figure available puts the gain from adding automatic slots alongside manual ones at over 5% of ad revenue on average. Moving breaks from bad positions to good ones is a change of a similar order: it shifts revenue per view by a few percent and protects some watch time you were losing.
Now compare that with the lever on the other side of the equation. Click-through rate on a browse-heavy channel routinely varies by a factor of two or three between a weak thumbnail and a strong one, and the number of views is what every revenue path in the previous paragraph is multiplied by. A few percent on monetisation of the views you already have is worth an afternoon. A better click-through rate is worth a working habit. If you have not established what normal looks like for your channel, what counts as a good CTR is the place to start.
A thirty-minute pass over your channel
- Open your three most-viewed uploads of the last ninety days and look at where the breaks sit against the retention curve. Anything in the first minute, inside a sponsor read, or over the end screen comes out now.
- On each of those three, leave automatic slots ticked and keep only the manual breaks you can justify at a structural seam. Check whether Studio flags any of them as interruptive.
- For the back catalogue, use the bulk path in Content to apply automatic mid-rolls to videos that have none. Confirm you have selected the option for videos without existing breaks rather than the one that overwrites everything.
- Check the monetisation status of anything with a suspiciously low RPM before touching its breaks. Limited ads are not a placement problem.
- Add one line to your editing checklist: a beat of silence at each section change. That is the whole of break design, done upstream.
- Close Studio and go back to the thumbnail and title, which is where the next order of magnitude lives.
Frequently asked questions
Do more mid-roll breaks mean more money?
Not proportionally, and not reliably. A second break earns only from viewers still present when it arrives, and heavy ad loads cost watch time that reduces the impressions the whole video receives. The arrangement YouTube's own data supports is manual breaks at genuine seams with automatic slots filling in — not the maximum number of slots you can fit.
Should I opt out of automatic ad slots?
Usually no. The one first-party number in this area points the other way, and automatic placement since May 2025 is aimed at natural break points. Opt out on the specific videos where you have a real reason — a sponsored segment you need protected, a tightly structured piece where every interruption is a creative decision — rather than as a channel-wide stance.
Can I put mid-rolls on a video shorter than eight minutes?
No. Below eight minutes a video can carry pre-roll and post-roll ads but no mid-roll breaks. The threshold has been eight minutes since 2020.
Does ad placement affect how the algorithm treats my video?
Not directly, as a setting. Indirectly, yes: interruptions that drive viewers away reduce watch time and satisfaction signals, and those do affect how much the system shows your video. The route runs through viewer behaviour, not through a penalty on ad settings.
Why do my older videos have more breaks than I remember setting?
Because of the May 2025 backfill. Videos uploaded before 24 February 2025 that had only manual mid-rolls received additional automatic slots at natural points. You can opt those videos out in Studio if you want your original placement restored.
Do Premium viewers see my mid-roll breaks?
No. Premium members watch without ads, and the revenue for their watch time comes from a share of subscription income distributed by how much they watch rather than from impressions.
Do Shorts have mid-roll breaks?
No. Shorts monetisation works through a pooled revenue share rather than ads placed inside an individual Short, so there is nothing to place. If short-form is the bulk of your output, your monetisation attention belongs on the pool mechanics rather than on break placement.
The short version
Ad formats stopped being your decision in 2023, and in 2026 the formats themselves got longer on the largest screen your audience uses. What is left to you is the moment of interruption, and the honest way to choose it is to stop looking at the timeline and look at the retention curve: a break is worth the audience still standing when it arrives, which is why two early breaks at real seams beat four spaced politely across a video nobody finishes.
Do the pass, tick the automatic slots, keep the manual breaks you can defend, and stop. The remaining upside in ad settings is a few percent, and it is bounded. Everything above the ad settings in the funnel is not.
Which is the uncomfortable part of any monetisation conversation. Revenue is views multiplied by how well those views are monetised, and the first term moves far more than the second. The breaks are a thirty-minute job; the thumbnail is the job. Thumblore generates thumbnail variants from a prompt or a frame in about a minute, which is what makes testing two genuinely different approaches cheap enough to do on every upload — and a video that earns twice the clicks carries every ad slot in it to twice as many people.