YouTube ads for performance, not branding.

YouTube advertising suffers from a positioning problem. Because it looks like television, it gets planned like television — reach, frequency, brand lift, a nice film and a prayer. Meanwhile a smaller group of advertisers runs it as what it has actually become: the largest pool of watchable direct-response inventory on the internet, with Google's conversion bidding underneath and intent data no TV buyer ever had. Both groups buy the same placements. Only one of them can tell you their cost per acquisition.

The performance toolkit

For conversion goals, your workhorses are skippable in-stream ads bought on conversion-focused bidding, and YouTube placements running inside Demand Gen campaigns — the campaign type that spans YouTube, Shorts, Discover, and Gmail with shared optimization. In-stream gives you the classic pre-roll moment with time to persuade; Shorts placements demand vertical creative and reward native-feeling content; in-feed catches people browsing. The strategic choice is between standalone video campaigns (tighter control, placement-level reporting, format isolation) and Demand Gen (broader reach, algorithmic allocation across surfaces). Our default: standalone for the first proving phase, Demand Gen once creative and conversion signal are strong enough to feed a broader system.

Audiences: intent, not interests

The most under-used asset in YouTube advertising is Google's search data. Custom segments built from the searches your buyers make — targeting people who recently searched your money keywords on Google, then showing them video — consistently outperform interest categories, which are too broad to carry direct response. The audience stack we deploy, in rough order of efficiency: remarketing and customer-list audiences first, custom segments from converting search terms second, in-market segments for the category third, and only then broader intent expansion once the account has conversion history for the algorithm to generalize from.

Creative: engineered for the skip button

Direct-response video is a different craft from brand film, and the skip button is the design constraint that changes everything. The structure that survives it: hook in the first five seconds (the problem, the pattern-interrupt, the outcome — not your logo), value demonstrated in the body with the product actually doing the thing, and a spoken, on-screen, unambiguous call to action. Make it work with sound off. Produce vertical for Shorts, not just letterboxed horizontal. And plan for volume — a performance channel needs a rotation of concepts and refresh cadence, the same testing discipline that runs on Meta, because fatigue on YouTube is slower than social but just as real.

Measurement: respect the lag

YouTube converts on a delay — people watch, then search you later, then convert on another device. Last-click will always under-credit it, sometimes to the point of recommending you kill a channel that is quietly filling the top of your funnel. The honest toolkit: include engaged-view conversions in evaluation (with a consistent policy, not opportunistically), watch branded search volume against video flight dates, and when spend justifies it, run a geo holdout — the same incrementality logic we apply everywhere. Judge the channel on blended account performance over 4-6 week windows, not on its own last-click line, or you will make the classic mistake: concluding YouTube "doesn't work" while your search campaigns harvest the demand it created.

YouTube rewards advertisers who bring direct-response discipline to a channel still priced for brand-budget tolerance. Start with remarketing and search-intent audiences, engineer creative for the skip, measure with lag respected — and scale only what survives honest measurement. When you are choosing between YouTube-only and the broader multi-surface play, the Demand Gen decision piece picks that apart, and our YouTube advertising service page shows how we run it end to end.