Mid-rolls need eight minutes. Everything after that is editorial: put breaks on cuts you already made, keep them out of the opening, and leave the end screen alone.
YouTube ad breaks come in three kinds — pre-roll, mid-roll and post-roll — and only mid-rolls have a length requirement: the video must reach eight minutes. Everything after that threshold is an editing decision, and the one that matters is landing each break on a cut the viewer was expecting anyway.
The eight-minute rule, and what it does to editing
A video under eight minutes gets pre-roll and post-roll only. At eight minutes, mid-roll becomes available, and YouTube documents the eligibility in its ad formats guidance.
That single threshold explains an enormous amount of what YouTube looks like. It is why so many videos run to eight minutes and ten seconds, and it is the most common reason a creator pads an edit.
Padding is a real trade and it should be made deliberately. A tight six-minute video stretched to eight is worse for every viewer and better for one revenue stream. A seven-minute video with a genuinely useful section cut for time is a different case entirely — put the section back. The rule is a fact, not an instruction.
Where YouTube ad breaks should go
Three constraints, in order of how much they matter.
On a cut. An ad between a question and its answer costs you the viewer. An ad at the boundary between two chapters is a pause the viewer had already been handed. If you write chapters, you have already done the work of finding these moments — the break just needs to be told where they are.
Not in the opening. Nobody has committed yet. A break in the first ninety seconds is the cheapest way to lose a viewer who was about to stay.
Not in the final minute. The end screen is the most valuable real estate on the video, and an ad immediately before it buys very little and costs the click through to your next video.
Between those guards, spacing is judgement. Every four minutes is aggressive; every six to eight is common; on a long-form video that people watch for the ending, fewer is usually the better trade.
Why automatic placement is not enough
YouTube will place mid-rolls for you, and its placement optimises for revenue rather than for your edit. It has no way to know that 8:42 is mid-sentence, because it did not cut the video.
Two minutes of manual placement is worth doing for a reason that is not about the ad at all: the cost of a badly placed break is not the thirty seconds of advertising, it is the viewer who does not come back after it. Retention loss compounds — it costs the watch time, the session, and the next recommendation.
The ad break planner takes the chapter list you already wrote for the description, snaps each planned slot to the nearest real boundary, refuses to place anything in the guards at either end, and hands back timestamps you can paste into Studio.
Plan a video's breaks
Enter the length as it reads in your timeline.
Paste your chapters exactly as they go in the description — 0:00 Intro per line.
Set the spacing you want. Six minutes is a sensible default.
Copy the timestamps and set them in Studio's ad break editor.
If the video is under eight minutes, the planner says so rather than handing you a list of mid-rolls YouTube will ignore. That is the honest answer, and it is one fewer thing to discover after publishing.
While you are working with timestamps, the timestamp link builder turns the same chapter list into deep links for a pinned comment or a description.
The mistakes that show up in the comments
Four placement errors account for most of the complaints a creator gets about ads, and none of them are about the number of breaks.
Mid-sentence. The most common, and always the result of automatic placement or of a timestamp copied from a rough cut rather than the final one.
Immediately after the hook. A break at 1:20, right after you have promised what the video will deliver and before you have delivered any of it, reads as a toll booth.
Two breaks close together. Usually an accident: a slot placed by hand and another left on automatic. Check that automatic placement is off if you are setting them yourself.
Just before the payoff. Structurally the worst one. If the video builds to a reveal, an ad in the ninety seconds before it converts anticipation into irritation, and that is the emotion the viewer carries into your end screen.
The planner's guards catch the first and the last of those by construction; the middle two are a matter of reading the list before you paste it.
How this shows up in your numbers
More mid-rolls raise the share of your views that carry an ad, which raises your RPM through the first term of the identity in CPM vs RPM. They also lower the share of viewers who reach the end.
Both effects are real, and the second one is invisible in the revenue report — you see the ads you gained and not the sessions you lost. The way to hold yourself honest is to watch average view duration alongside RPM after a change, and treat a drop in the first as a cost of the second.
How long does a video have to be for mid-roll ads?
Eight minutes. Below that, only pre-roll and post-roll are available, regardless of the video's watch time or the channel's size.
How many ad breaks should I run?
Fewer than the maximum. Every four minutes is aggressive; six to eight is common. On a video people watch for the ending, extra breaks usually cost more than they earn.
Should I let YouTube place the breaks automatically?
Automatic placement optimises for revenue, not for your edit, so it will land an ad mid-sentence. Placing them by hand takes two minutes.
Do ad breaks hurt watch time?
They can. Watch average view duration alongside RPM after any change to the ad load — the revenue gain is visible and the retention cost is not.
Price on reach times a niche rate times your engagement band — never on followers alone. Quote three numbers: the ask, the fair rate, and the floor you walk away below.
YouTube publishes the categories that limit ads. The opening of a video is weighted hardest, context decides everything, and no word list can see context.