YouTube ads

The YouTube ad formats and what each one buys

Skippable in-stream ads play before or during a video with a skip button after five seconds. You pay when someone watches thirty seconds, watches to the end, or interacts. That pricing model is generous: people who skip cost you nothing, so the format rewards a strong opening rather than punishing a long ad.

Non-skippable in-stream ads run fifteen to twenty seconds with no escape, and you pay per thousand impressions. They guarantee the full message but earn no goodwill, so they suit short, single-idea messages where reach is the point. Bumper ads are six seconds, unskippable, and best used for one memorable idea repeated across a campaign.

In-feed video ads appear in search results and alongside related videos, so the viewer chooses to click. They behave more like search advertising and suit content people would actively look for. YouTube Shorts ads sit in the vertical feed and follow short-form rules rather than YouTube ones.

• Skippable in-stream: the default for consideration and action
• Non-skippable: guaranteed message, best kept short and simple
• Bumper: six seconds, one idea, good for frequency and recall
• In-feed: the viewer opts in, so treat it like search intent

Writing for the first five seconds

Assume the skip. Every viewer has their cursor over the button, so the opening has to earn the next five seconds rather than build towards a reveal. Lead with the thing that matters: the problem, the offer or the surprising claim. Save the brand film structure for people who chose to watch.

Show the brand early. This runs against the instinct to hold it back for the end, but if most viewers skip at five seconds, an end card reaches almost nobody. Brand recognition in the opening seconds is one of the few things that measurably survives a skip.

Then reward the people who stay. The ad should get more useful as it goes, not more repetitive. A well-built thirty second ad has a hook, a demonstration and one clear ask. Longer ads can work when the content genuinely holds, and on skippable inventory you only pay for people who chose to keep watching.

Audience targeting: intent beats interest

YouTube inherits Google's audience data, which makes intent-based targeting its real strength. Custom segments built from search terms let you reach people who recently searched Google for what you sell, while they watch YouTube. That is closer to search advertising than to display, and it usually outperforms broad interest categories.

In-market segments identify people Google believes are actively researching a purchase category. Life event targeting catches moves, marriages and job changes. Detailed demographics cover things like parental status and homeownership. Affinity audiences are broad interest groups and suit reach campaigns more than conversion ones.

Placement targeting deserves a mention because it is unique to this platform. You can choose specific channels, specific videos or topic groups. It is useful for reaching a defined enthusiast audience, and equally useful in reverse: excluding made-for-kids content, low-quality channels and categories that do not fit the brand.

• Custom segments from recent search terms: the highest intent option
• In-market segments for active category researchers
• Your own remarketing lists, usually the best performing audience of all
• Placement exclusions to keep the brand away from unsuitable content

Video action campaigns and driving conversions

Video action campaigns are the format for conversions. They combine in-stream and in-feed inventory, add a persistent call to action and headline, and bid towards conversions rather than views. If the goal is enquiries or sales rather than awareness, this is where to start.

They need conversion data to work, which is the usual condition on any automated bidding. An account with no conversion history will get better results buying views cheaply while it builds the signal, then switching to action bidding once there is something to optimise against.

Feeds change what these campaigns can do. Connecting a product feed turns the ad into a browsable set of products beneath the video, which suits retail well. For lead generation, adding lead forms directly to the ad removes a step, though as everywhere the easier form brings some lower-intent enquiries with it.

Measuring YouTube honestly

Judging YouTube on last-click conversions alone will usually make it look worse than it is. Video creates demand that surfaces later through search, direct visits or a return trip weeks afterwards. Some of that is captured by view-through conversions, which count people who saw the ad, did not click, and converted within the window.

View-through conversions are useful but easy to over-trust, since some of those people would have bought anyway. Treat them as directional evidence rather than as a number to bill against. Brand lift studies give a cleaner read on awareness and consideration, though they need a certain spend level to run.

The most practical measure for smaller advertisers is simpler. Watch branded search volume and direct traffic while a YouTube campaign runs, and again when it pauses. If the campaign is doing its job, both move. That is a blunt instrument, but it answers the question that matters: is this spend producing demand elsewhere in the business?

• View-through conversions: useful signal, not a number to bank
• Watch branded search and direct traffic through the campaign period
• Test with pauses, the drop tells you more than the dashboard
• Brand lift studies where the budget supports them

Sequencing and frequency across a campaign

Ad sequencing lets you show videos in a set order to the same person. A short hook first, a demonstration second, an offer third. It maps neatly onto how people actually decide, and it avoids the awkwardness of asking a stranger to buy in the first exposure.

It works best where the product needs some explaining and where the audience is defined enough to build the sequence around. It works badly on very broad audiences, where most viewers only ever see the first step and the later videos never get their run.

Frequency management matters on YouTube because the same viewer can be served your ad repeatedly across a viewing session. Set frequency caps on reach and awareness campaigns. On action campaigns the bidding usually handles it, but it is still worth checking the frequency report, because an ad seen eight times in a week is building irritation rather than familiarity.