Demand Gen is the campaign type most Google Ads accounts either ignore or run with Search logic, and both are mistakes. It replaced Discovery in 2024, absorbed YouTube into a single campaign type, and operates on a different premise from everything else in the account: it creates demand instead of harvesting it. Here is where it fits, what to feed it, and how to judge it honestly.
What Demand Gen is and how it differs from PMax and Search
Demand Gen is Google's visual, audience-driven campaign type, serving image and video ads across YouTube, Discover, and Gmail. Search is keyword-driven: someone asks, you answer. Performance Max is goal-driven: you hand Google assets and a target, and it buys across every surface it owns with limited visibility into where.
Demand Gen sits between them. You choose the audiences, control the creative per ad, and can see where it runs, which makes it the most controllable of Google's visual formats by a wide margin. That control is the reason it deserves its own playbook instead of inheriting one from Search.
If the PMax versus Search distinction is about two flavours of demand capture, Demand Gen is a different job description entirely. The same principle from our PMax versus Search breakdown applies here: pick the tool by the job it does, not by what Google is promoting this quarter.
Where Demand Gen fits in your Google Ads funnel
Search and Shopping harvest demand that already exists. Their ceiling is search volume: once every relevant query is covered, more budget buys more expensive clicks, not more customers. That ceiling is where most accounts hit diminishing returns, the pattern we documented in why more ad spend stops working.
Demand Gen exists for the people who are not searching yet. They match your customer profile, they scroll Discover and YouTube daily, and they have never heard of you. The campaign's job is introduction, not closing, and briefing it as a closer guarantees disappointment.
Judge it as a pipeline builder. It feeds the audiences and the brand searches that your capture campaigns convert later, which is exactly why measuring it like a closer makes a working campaign look broken.
Placements: what YouTube, Discover, and Gmail each deliver
Three surfaces, three different qualities of attention. YouTube delivers in-stream, Shorts, and in-feed placements. It is the strongest format for showing a product in use, and the one worth designing creative for first.
Discover is the scrolling feed on the Google app and Android home screens. It behaves like native social inventory: visual products with strong lifestyle imagery earn cheap, engaged sessions, while flat studio shots get scrolled past without registering.
Gmail places ads in the Promotions and Social tabs. It is usually the cheapest reach in the campaign and the lightest attention, fine as bundled inventory, weak as a primary placement. Check the per-surface split in reporting once data accrues, and let what you find shape the next round of creative.
Creative requirements: the asset coverage that determines reach
Asset coverage decides reach. Demand Gen wants landscape, square, and vertical versions of your images, video in the same three ratios where possible, plus the standard stack of headlines and descriptions. Missing ratios means ineligible inventory, and thin asset groups leave the algorithm nothing to rotate when an ad fatigues.
In feed environments, the creative is the targeting. The image does the qualifying that keywords do in Search, so product-in-context lifestyle visuals consistently beat white-background catalogue shots in the accounts we manage.
Plan refresh from day one. Feed creative wears out in weeks, not quarters, and a fatigued Demand Gen campaign does not fail loudly. It just quietly degrades into expensive wallpaper while the reports stay green.
Audience strategy: who to target and how to layer remarketing
Start where the signal is strongest. Customer match lists and site visitors give the algorithm a seed of people who already buy from you. Lookalike segments built from those lists extend reach to people who resemble them. In-market and affinity segments come last, as the coldest layer, with creative doing the qualifying.
Layer remarketing deliberately rather than letting it blur into prospecting. A returning cart abandoner and a complete stranger deserve different creative, different messaging, and different expectations in your reporting.
The upstream effect is real but indirect. On our fashion account running at 17.3x ROAS, visual brand presence upstream contributed to the lower-funnel performance the account gets measured on. Channels compound; attribution rarely shows it cleanly.
Measurement: why last-click undersells Demand Gen
Last-click attribution gives Demand Gen almost nothing. A shopper sees your product on YouTube, searches your brand three days later, clicks the brand ad, and Search collects the conversion. Read literally, the report says Demand Gen failed during the exact weeks it did the introducing.
Better signals exist: growth in brand search volume after launch, new-customer rate on the conversions the campaign touches, view-through and assisted figures treated as direction rather than gospel, and geo holdout tests once spend is large enough to justify them.
Set the measurement framework before launch and agree what success looks like at week six. Deciding how to judge the channel after the data arrives is how demand creation gets killed prematurely in almost every account that abandons it.
Budget allocation: sizing Demand Gen against Shopping and Search
Capture comes first. If Search, Shopping, and your product feed are not converting efficiently, Demand Gen will pour water into the top of a funnel that leaks at the bottom. Fix capture, then fund creation.
For a starting size, a modest single-digit share of total Google Ads spend is enough to read signal without betting the quarter, scaled up as brand search and new-customer numbers respond. Treat the budget as a test with a review date, not a permanent line item, and hold it to the measurement framework you agreed before launch.
If you want a straight answer on whether your account has the capture foundation to justify it, the free 48-hour audit answers exactly that.