YouTube Bumper Ads: The 6-Second Format, End to End

YouTube bumper ads are the format I reach for when the brief is "make sure a lot of people remember one thing." Six seconds, unskippable, priced on impressions — there is nowhere to hide and nothing to over-engineer. After years of running these across brand budgets, my strongest opinion is that most advertisers misread them: they judge a bumper by clicks and bounce off the format, when the format was never built to be clicked.

This article is the end-to-end version. What the format is, the specs, what it costs, the benchmarks I actually see, how to build one, and — the part that matters most — when bumper ads earn their place in a plan versus when you're using them wrong. I'll flag where my experience diverges from the tidy answer, because the tidy answer is where a lot of wasted spend hides.

What are bumper ads on YouTube?

Here's the plain definition first, then the inference. Bumper ads on YouTube are a short in-stream video format: six seconds or less, non-skippable, playing before, during, or after another video. The viewer cannot skip, and you don't need them to click — you pay to be seen, not to be acted on.

Now the part that changes how you use them. A bumper is a frequency instrument disguised as a creative unit. Because it's so short and so cheap per impression, its job in a media plan is to land a single association again and again until it sticks. That's a completely different animal from a skippable in-stream ad, where the viewer opting to stay is the whole signal. If you want the side-by-side on that distinction, my breakdown of bumpers versus TrueView is the piece I'd read next — the choice between them is really a choice between buying frequency and buying self-selected attention.

The mental model I use: a bumper is a billboard you can't drive past. Six seconds of guaranteed exposure, no exit. You wouldn't put a paragraph on a billboard, and you shouldn't put a pitch in a bumper. One image, one line, one brand cue — that's the whole canvas.

Bumper ad specs and what they cost

The specs are refreshingly simple, which is part of why I like the format for teams new to video. Your video must be hosted on YouTube (public or unlisted), it must be six seconds or shorter, and the ad serves as non-skippable in-stream. Google Ads assigns the bumper format automatically off the duration — feed it a 6-second cut and you get a bumper; feed it 8 seconds and you've quietly built a standard non-skippable ad instead. Horizontal 16:9 is the safe default, and I always supply a companion banner where the placement supports it.

How bumper ads bill

This is the load-bearing spec. Bumper ads run on Target CPM bidding, so you pay per thousand impressions, not per click and not per view. That single fact should reset your expectations: you are buying reach, and the cost of an impression is the number to watch, not cost-per-click. For the authoritative, always-current description of how the format serves and bills, Google's own about video ad formats page is the reference I keep open next to my own numbers.

Cost and benchmarks I actually see

Because you're bidding CPM, the useful benchmark is the effective cost per thousand, and in the accounts I run bumper CPMs typically sit in a low single-digit to low double-digit range depending on country, audience tightness, and competition. That makes bumpers one of the cheapest ways to buy broad reach on YouTube — they're consistently cheaper per impression than a longer non-skippable spot. If you want the wider picture on what video inventory costs and how the CPM-versus-CPV pricing models compare, I've laid that out in the guide to what YouTube advertising costs. The number I judge a bumper campaign on isn't CPC at all — it's CPM against reach and, where I can measure it, brand lift.

When I use bumpers for reach and frequency — not conversion

This is the section that saves budgets, so I'll be blunt about it. Bumper ads are a reach-and-frequency tool. I reach for them when the objective is awareness at scale — a launch, a seasonal push, a brand I need front-of-mind before a demand moment — and I stop reaching for them the second someone frames the goal as "get conversions." The format bills on impressions and gives the viewer nowhere to click; asking it to be a performance channel is asking a billboard to close a sale.

The inference from how the format is priced is clean: Google built bumpers around CPM because the value being sold is memory, not action. So the right KPIs are reach, frequency, and lift, and the right way to buy them is inside a reach-oriented objective where the system optimizes for unique users at your target CPM. When I want the same person hit several times to build recall, I structure the campaign around a frequency goal and let bumpers carry the repetition — short enough that the repeated exposure doesn't grate the way a repeated 30-second spot would.

Where bumpers genuinely earn their keep is as the finisher on a longer story. I'll run a skippable in-stream ad to do the explaining and heavy lifting, then follow with bumpers to re-stamp the one line I need remembered. That sequencing — long form to teach, bumper to remind — is where I see the reach investment convert into recall. The mechanics of the reach and frequency objectives, and which bidding subtype to pick, live in my rundown of video campaign bidding strategies; the choice you make there is what actually decides whether your bumpers buy fresh reach or repeated frequency.

How to set up a bumper ad campaign

The setup is short, and the only place people go wrong is the bid strategy, so I'll flag that step hard.

The build, step by step

Start a new Video campaign and choose an awareness or reach goal — efficient reach or non-skippable reach both make bumpers available, and if you want repeated exposure, pick the frequency-oriented subtype. Set your daily budget, locations, and languages, then choose your audiences. Under bidding, make sure you're on Target CPM; this is the switch that lets the bumper format serve at all. Then paste in your YouTube-hosted video of six seconds or less, and Google Ads will recognize it as a bumper automatically. Add a companion banner and a final URL so the desktop click has somewhere to land, and launch.

One operational note from experience: keep bumper ads in their own campaign or ad group rather than mixing them with longer video. It keeps the reporting honest — you can segment by ad format, but a clean split makes it far easier to read bumper CPM and reach without a 30-second spot muddying the averages. If you're building a broader video plan, the wider format landscape and where bumpers sit within it is covered in the full catalogue of YouTube ad formats, which is the pillar I'd anchor the plan to.

Bumper ad best practices that actually move the needle

Six seconds sounds trivial until you try to fit a brand into it. Here's what consistently works in the accounts I run.

One idea, no more. A bumper that tries to say two things says nothing. Pick the single association — a product, a category cue, a tagline — and build the whole six seconds around it. The best bumpers I've run could be described in one sentence.

Brand early, not at the end. Because there's no skip, viewers do watch, but attention still front-loads. Get the logo or product on screen in the first two seconds; a reveal that saves the brand for second six is a reveal a lot of people mentally checked out before reaching.

Design for sound-off, reward sound-on. Assume muted playback and make the visual carry the message, then let audio add a layer for the people who have it on. A bumper that only works with sound is a bumper that fails half its impressions.

Run bumpers in flights, not forever. Their power is frequency inside a window — a launch, a sale, a campaign burst. Left running indefinitely at high frequency, the same six seconds starts to wear. I set frequency targets and refresh the creative before fatigue sets in rather than letting one cut run down.

Judge them on the right scoreboard. Reach, frequency, CPM, and brand lift — not CTR, not conversions. Every time I've seen a team abandon bumpers as "not working," they were grading a reach format on a performance report. Set the KPI before you launch and the format looks a lot smarter.