Skip to content

Core mechanics · module 8 of 13 · 7 min read

Creative and assets

When targeting is automated, what is left to control?

  • responsive search ads
  • asset groups
  • asset quality
  • ad strength
  • creative testing

Read first: Campaign types and where ads appear

As targeting and bidding automate, creative becomes the largest remaining lever — and it is the one most accounts under-invest in, because it is not a setting you can toggle.

Responsive formats

Responsive search ads take multiple headlines and descriptions and assemble combinations per auction. You supply the range; the system picks. Practical implications: give genuinely different angles rather than fifteen paraphrases, and use pinning sparingly — pin only what must appear, such as legally required wording.

Asset groups in Performance Max work the same way across every channel: text, images, video, logos. If you do not supply video, Google generates one, and auto-generated video is usually worse than anything you would ship deliberately.

Ad strength is a feedback tool about diversity and coverage, not a performance prediction. Do not chase "Excellent" at the cost of message quality.

Google now publishes the numbers

In August 2026 Google released YouTube creative guidance with research attached: Nielsen work attributing 49% of campaign ROI to creative, and Ekimetrics research finding that improving creative can more than double ROI on YouTube.

The specifics for Demand Gen in-stream skippable ads (US, July 2026 data):

  • including a human voice — about 12% higher conversions on average
  • adding supers or a text overlay — about 3% higher
  • visualising the brand early in the video — also associated with lifts

Google also advises removing fake buttons from image assets and making creative look native to the platform rather than like a repurposed TV spot.

If creative is roughly half of ROI, it deserves roughly half the optimisation effort. Almost no account allocates time that way.

Testing creative honestly

Creative tests need enough conversions to separate signal from noise, and most accounts do not have the volume for weekly winners. Test big differences — a different offer, a different angle — not button colours. Give tests time, and accept that many will be inconclusive.

Landing pages

The ad's job is to earn a click; the landing page's job is everything after. Landing page experience is also part of auction-time quality, so a poor page raises costs as well as lowering conversion rate. Message match matters: an ad promising a specific thing must arrive somewhere that delivers that thing above the fold.

What trips people up

  • Fifteen near-identical headlines and calling it a responsive ad.
  • Judging creative on CTR rather than downstream value.
  • Never refreshing, then blaming the algorithm for fatigue.

You have got this when

Your first suggestion for an underperforming campaign with sound structure and tracking is a creative and offer change, not a bid adjustment.

Go to the source

What has changed since

Stories from the briefs that touch this module.