Pre-Roll, Mid-Roll, Post-Roll: Where YouTube Ads Land

Almost every question I get about the YouTube ad formats pre-roll mid-roll post-roll trio starts from the wrong place: people treat them as three different kinds of ad. They aren't. They're three positions for the same in-stream ad — before the video, partway through it, and after it ends. The creative can be identical; what changes is the moment you catch the viewer, and that moment does more to your finishing rate than any tweak to the ad itself.

I've run the same fifteen-second spot in two of these slots on the same audience in the same week, and watched the numbers diverge in ways the creative alone can't explain. This page branches off the main map of every YouTube format: "in-stream" quietly hides three viewer contexts, and once you see them as positions, the pre-roll mid-roll post-roll youtube question stops being about naming and starts being about timing — a bet on where the viewer's attention already is when your ad arrives.

The three slots, before during and after

Google's in-stream ads are documented as playing before, during, or after other videos — those three words are the positions, and the difference is entirely about the viewer's state when the ad lands.

Pre-roll (before the video)

Attention is high but impatient — they're leaning toward the thing they clicked, not toward you, and if it's skippable the five-second countdown is the most-watched real estate on YouTube. The workhorse slot: highest reach, and the one place a strong hook pays for itself instantly. The tax is that anticipation is aimed elsewhere, so a slow open dies here fast.

Mid-roll (during the video)

Interrupts something the viewer is already invested in, on videos long enough to carry a break. Counterintuitively, this is often my best finishing slot: the viewer is committed to getting through, so they tolerate the ad as the toll for getting back to it. The sunk cost of the half-watched video works in its favor. The catch is inventory — it only exists on longer runtimes, so reach is narrower.

Post-roll (after the video)

Plays once the content is over, and it's the weakest of the three: the viewer got what they came for and is already reaching for the next clip. I use it sparingly, almost never as a primary slot — only for cheap incremental frequency against a warm audience that's already seen the real message.

For how skippable in-stream charges before you layer position on top, the piece on how skippable ads bill and serve is the natural prerequisite here.

How attention and completion shift by position

This is the part the spec sheets never explain, and it's the whole game. The same ad in three positions gets three different finishing curves.

Pre-roll maximizes exposure but fights impatience. A large chunk of viewers are counting down to skip, so effective attention is front-loaded into the first few seconds. I read pre-roll almost entirely through the opening hook — if the first five seconds don't hold, the rest is a rounding error.

Mid-roll trades reach for depth. The viewers you reach are in a finish-friendly state, so view-through tends to climb even as absolute impressions fall. Inference from my own accounts, not a Google-published number: the invested viewer finishes higher because they want their video back.

Post-roll is cheap but weak for cold objectives. Against a remarketing pool it's a low-cost reminder — reinforcement, never a slot that has to carry comprehension. Expecting it to convert a stranger is the classic error.

The practical upshot: I don't read the completion rate as a fixed property of the ad. I read it as the ad times the position — a "low" pre-roll and a "high" mid-roll can be the exact same file. To see how those choices ripple into what you pay, my breakdown of what YouTube ads cost across formats lays out how the billing interacts with where the ad lands.

Where overlay ads fit — and why they no longer do

People still list overlay alongside the three in-stream positions, so it's worth being clear: overlay was a fourth position — not before, during, or after, but on top of. It was the semi-transparent banner that slid across the lower third of a playing video, a display unit layered over the content rather than an in-stream break.

Here's the thing to internalize for 2026: overlay has been discontinued. Google retired the format, and it's no longer something you can build. If it shows up in an old template, treat that template as out of date — the same way "TrueView" as a format name and the old Discovery ads are gone. Budgeting for a slot that doesn't exist wastes a planning cycle. The current picture is three in-stream positions: before, during, after.

Position crossed with skippable vs. non-skippable: the real grid

This is the angle most guides miss. Position isn't a single choice — it's one axis. The other is whether the ad is skippable, and the two multiply into a grid that predicts behavior. You're not choosing pre-roll or skippable; you're choosing a cell.

Skippable × pre-roll is the highest-volume, most forgiving cell, where most of my direct-response and consideration spend lives. The skip button qualifies the audience for me: uninterested viewers leave for free, and I only pay for the ones the hook held.

Non-skippable × pre-roll guarantees the whole short message before the content, at the cost of goodwill. I use it for awareness where the full view genuinely matters and the message is tight.

Skippable × mid-roll is my quiet favorite for finish-sensitive work: the invested viewer is unlikely to skip, so I get near-forced view-through without forcing it. Non-skippable × mid-roll stacks a guaranteed full view onto that committed state — powerful, but the most intrusive cell, so I reserve it.

Post-roll, either way, is the reinforcement corner. I almost never run it non-skippable; the skippable version as a cheap warm-audience reminder is the only one I keep on. The point of the grid: "which position" and "skippable or not" are the same decision made twice, and the good cell depends on the funnel job.

When I pick what: the routing I actually run

Here's the decision I make before opening the campaign builder, because the position and skip mode have to match the job the funnel stage is asking for.

For cold reach and awareness, I default to skippable pre-roll for scale, and reach for non-skippable pre-roll only when the whole short message must be seen — I'm buying breadth, not depth.

For consideration, where I need the viewer to absorb a demonstration, I lean into skippable mid-roll on longer runtimes. That committed-viewer state is the single best finishing environment I have, and it lets a longer ad breathe without the pre-roll skip cliff.

For lower-funnel frequency against warm and remarketing pools, I mix skippable pre-roll with cheap skippable post-roll reminders — reinforcement doesn't need the attention that acquisition does.

The mismatch I see most often, across audits of accounts I inherit, is a post-roll or non-skippable slot being asked to introduce a brand to a cold audience — the position structurally can't carry that job. Since positions increasingly get assembled for you by automated campaign types rather than hand-picked, my thinking has shifted from "which slot do I select" to "which slots do I want the system leaning toward, and does my creative survive each one." For the always-current reference on how each position serves, bills, and combines with campaign types, Google's own about video ad serving reference is the page I check whenever a detail shifts.

The through-line: these three slots aren't a menu of styles, they're three moments in the viewer's attention, each crossed with a skip decision. Read them as positions, match the moment to the funnel job, and finishing stops being a surprise and starts being something you chose.