nova Creative Affinity: Google Demand Gen

Creative Affinity now analyzes Google Demand Gen alongside Meta, TikTok, and YouTube — so the visual, feed-based creative you're running across YouTube Shorts, Discover, and Gmail sits in the same view as everything else.
​

This matters because Demand Gen is where a lot of mid-funnel creative lives. It's the channel that warms people up before they ever search for the brand, and until now that creative was invisible in your omnichannel creative analysis. You can now see which Demand Gen assets are genuinely resonating, which ones are bringing in high-value customers, and how they compare to the same or similar creative running on Meta and TikTok.
​

Demand Gen is Google's feed-based campaign type. It runs image, carousel, and video ads across YouTube Shorts and In-Feed, Discover, Gmail, and Display inventory, and it targets using Google's AI and audience signals rather than keywords.
​

It sits between Display (pure reach) and Search (pure intent), aimed at mid-funnel consideration — the "I didn't know I wanted this yet" moment.
​

What it's good at

How clients typically use it

Demand Gen data is available at all three levels, same as your other platforms:

Metrics available at every level:

Demand Gen ads come through in their native formats — multi-asset, video responsive, and carousel — so you can compare a carousel against a video against a static within the same channel.
​

A few things to keep in mind when you're interpreting the numbers, so you draw the right conclusion:
​

Creative variety is the biggest performance lever. Demand Gen algorithmically tests combinations across formats and surfaces. An asset group with one image and one headline will underperform an asset group with real variety, almost regardless of how good that single asset is. If a Demand Gen campaign is flat, check asset depth before you blame the creative itself.
​

Judge it on assists, not last click. Like Display, Demand Gen is assist-heavy. Last-click ROAS alone will undersell it. Look at View-through Conversions and Affinity Score alongside ROAS, not instead of them.
​

Audience signal quality drives results more than exclusions. Because targeting is AI-driven off signals rather than manual placements, the quality of the seed audience or customer list matters more here than it does on Display.
​

Compare it to Meta, not to Search. Demand Gen's closest behavioral cousin in your creative set is Meta — broad AI targeting, visual feed placement, mid-funnel intent. That's the comparison that produces useful creative learnings. Comparing Demand Gen CVR to Search CVR will only tell you that Search captures intent.
​

Affinity Score works the same way it does everywhere else — a blended score built from engagement, clickthrough, and conversion signals that tells you what's resonating, fast. It's available for ads with at least $25 in spend and refreshes with every data load. Use it to spot promising Demand Gen creative early, before revenue rolls in.
​

LTV follows the exact same process as Meta, TikTok, and YouTube. It's based on matching a user in GA4 to a real transaction in the client's e-commerce platform, then attributing that customer's lifetime revenue back to the creative they engaged with. LTV is static once calculated and isn't affected by the date filter.
​

As always: Affinity Score is your early indicator, LTV is your confirmation. A Demand Gen ad with a high Affinity Score and high LTV is doing exactly what mid-funnel creative is supposed to do.
​

You can also drop Demand Gen creative into the Compare view (up to 4 creatives side by side) and the Creative Aggregations view, to see how a single asset performs across different campaigns and placements.
​

The more integrations in place, the better the matching algorithm performs, and the more accurate the LTV and CAC figures attached to each creative.
​

Why does the same creative perform differently on Demand Gen than on Meta?

Demand Gen runs one campaign across YouTube Shorts, Discover, Gmail, and Display at the same time, assembling assets into different combinations per surface. The same video can win on one surface and lose on another. Use the Creative Aggregations view to see where an asset was actually served before concluding it works less well on Google than on Meta.
​

My Demand Gen numbers look flat. Is it the creative?

Check asset depth first. Demand Gen tests combinations across formats and surfaces, so an asset group built on one image and one headline will underperform a varied one almost regardless of how strong that single asset is. Thin asset sets are the most common cause of flat Demand Gen performance.
​

Should I compare Demand Gen against Search?

Not on CVR or CTR — Search captures existing intent and will always win those. Demand Gen's closest behavioral match in your creative set is Meta, so that's the comparison that produces useful creative learnings. Across channels, compare on Affinity Score and LTV instead.
​

Affinity Score is high but LTV is low. What does that mean?

The creative is doing its job as mid-funnel content — it's compelling and driving engagement — but it hasn't led to tracked purchases yet. For Demand Gen specifically, that's often expected rather than a problem. Check whether the campaign is genuinely positioned as a consideration play, and look at what Search and Meta are converting downstream of it.
​
​Looking for something else?

Questions about integrations, data load timing, comparing across platforms, and which metrics the date filter affects are answered in the nova Creative Affinity: Google Ads overview article.