Advertisers often talk about "running YouTube ads" as if it's one thing, but the formats behave like different media channels wearing the same badge. A skippable in-stream ad, a 15-second non-skippable bumper, and a Shorts feed ad have different cost models, different viewer attention patterns, and different jobs to do in a funnel. Picking the wrong format for the campaign goal is one of the most common and avoidable sources of wasted YouTube spend.

Core YouTube ad formats

Format	Length / Skip	Billing model	Best funnel stage
Skippable in-stream	Any length, skippable after 5s	CPV (counted view) or CPM, depending on bid strategy	Consideration — rewards ads that earn attention past 5s
Non-skippable in-stream	15-20s, cannot skip	CPM, guaranteed full view	Awareness / brand reach where forced exposure matters
Bumper ads	6s max, non-skippable	CPM	Awareness, frequency/reminder layer on top of other formats
In-feed video ads	Thumbnail + text, click to watch	CPC (on click to view)	Consideration — relies on thumbnail/title pull, not forced view
Shorts ads	Vertical, full-screen, native feed	CPM, part of Video Reach/Demand Gen campaigns	Awareness and younger-skewing reach, short-form attention pattern

Skippable in-stream: the 5-second test

Skippable ads only bill (under CPV bidding) when a viewer watches 30 seconds, watches to completion if shorter than 30 seconds, or interacts with the ad — a skip within the first 5 seconds costs nothing. This changes how creative should be built: the first 5 seconds carry disproportionate weight, since an ad that doesn't hook in that window gets skipped and generates impression without cost, but also without any brand exposure worth counting. Front-loading the brand or product visual in the first 5 seconds, rather than a slow build-up, is the standard practical response.

youtube_format_selection.txt

text

Goal: Drive demo signups (consideration)
  -> Skippable in-stream, CPV bidding, hook in first 5s
  -> Pair with in-feed ad using a curiosity-driven thumbnail

Goal: Build brand reach for a launch (awareness)
  -> Non-skippable 15s or bumper ads, CPM, high frequency cap
  -> Layer Shorts ads for younger/mobile-first reach

Goal: Re-engage past visitors (retargeting)
  -> Skippable in-stream to a remarketing list audience
  -> Shorter cut (15-30s) of the main creative, direct CTA overlay

View-through conversions: the format-specific measurement wrinkle

YouTube attributes two distinct conversion paths: a view-through conversion (VTC) counts when a user sees (but doesn't click) an ad and converts within the attribution window without any further ad interaction, versus a click-through conversion, which follows an actual click. VTCs are counted differently depending on format — for skippable in-stream, a 'view' that counts toward VTC eligibility generally requires the viewer to have watched a meaningful portion of the ad (not just had it technically load), while for non-skippable and bumper formats, the full forced exposure itself qualifies as the counted view.

This matters because VTC windows are generous by default (commonly 3 days, configurable), and VTCs get folded into the same 'conversions' column as click-throughs in Google Ads reporting unless you explicitly segment by conversion path. An account running heavy brand-awareness, non-skippable or bumper inventory will show a conversion count that is substantially inflated by VTCs relative to an account running primarily skippable in-stream or in-feed, simply because more people see the ad to completion without ever intending to click.

Don't blend VTC and click-through conversions in ROAS reporting

A view-through conversion means someone saw the ad and later converted through some other path (organic search, direct visit, another channel) — it is correlational, not causal, evidence of ad impact in the way a click-through conversion is. Reporting blended conversion value across both paths without segmentation routinely overstates YouTube's efficiency relative to click-based channels, because the VTC pool captures conversions that may have happened anyway. Always break out the segment column by conversion path before comparing YouTube ROAS to a Search or Shopping campaign.

Shorts ads borrow Demand Gen's measurement, not classic in-stream's

Shorts ads run through Video Reach campaigns or Demand Gen campaigns rather than classic in-stream video campaigns, which means they inherit Demand Gen's attribution and reporting conventions (engagement-based signals, feed-native metrics) rather than the VTC/CTC split described above for in-stream. Don't assume Shorts performance data lines up cell-for-cell with in-stream reporting when comparing formats.

Frequency and sequencing across formats

Because bumpers, non-skippable, and skippable ads can run in the same campaign group with frequency capping applied across them, a common sequencing pattern is to show a longer skippable ad first (to build fuller understanding) and use bumper ads as a cheap, high-frequency reminder layer afterward. YouTube's native Ad Sequencing feature supports this directly — defining a step 1 / step 2 / step 3 creative order a given viewer moves through, rather than randomizing which ad an eligible viewer sees.

What's next

Video isn't the only channel where attribution gets fuzzy once you leave click-based measurement — podcast advertising faces an even sharper version of the same problem, since there's no view or impression pixel at all to lean on.

Next: Podcast Advertising →

I build these systems professionally.

Whether it's a RAG pipeline, analytics migration, or AI workflow — let's talk.

Work With Me →

Need production AI or MarTech expertise? I build custom LLM workflows, server-side GTM tagging, and analytics infrastructures professionally.

Need custom AI or MarTech setup? Let's build together.

Let's talk →