Demand Gen vs Meta Ads: Two Platforms, One Playbook?

For most of Google's existence as an advertising platform, the product was fundamentally about precision. Keywords. Intent. Someone types a query; you show an ad. The tighter the match, the better the result. That's what made Google Ads different from everything else — you weren't interrupting people; you were meeting them exactly where they already were.

That era hasn't ended, but it's been joined by something quite different. Exact match isn't what it used to be. Broad match now powers much of Smart Bidding's audience-finding behaviour. Performance Max runs across every Google channel from a single campaign with minimal targeting control. And in 2023, Google launched Demand Gen — a campaign type explicitly designed for top-of-funnel, running on YouTube, Discover, and Gmail with creative-driven, audience-signal-based targeting.

In other words: Google has built a product that works like Meta. And if you're running both channels without recognising that convergence, you're probably making strategic mistakes on both.

What Demand Gen actually is

Demand Gen is Google's top-of-funnel campaign type, launched to replace the older Discovery campaign format. It serves ads across YouTube In-Feed, YouTube Shorts, Discover, and Gmail — reaching users who aren't actively searching for anything. It's interruptive by design.

Like Performance Max, Demand Gen relies heavily on Google's machine learning to find audiences. You provide audience signals — interest-based audiences, customer lists, lookalikes — but the algorithm has significant latitude in deciding who actually sees your ads. You provide creative assets (images, video, headlines, descriptions), and the algorithm assembles and optimises combinations.

This is structurally different from Search. In Search, targeting is explicit — you define the keywords, the algorithm finds the searchers. In Demand Gen, targeting is signals-based — you provide the inputs, the algorithm finds the audience. The quality of your creative and the quality of your audience signals determine how well it works.

What Meta Ads actually is (now)

Meta Ads has always been interruption marketing — you appear in a feed, users didn't ask for you, you need to earn their attention. What's changed significantly since iOS14 is how targeting works.

Before iOS14, Meta's targeting was genuinely precise. Interest layers, demographic combinations, behavioural segments — you could construct detailed audience definitions and be reasonably confident the platform was reaching them. That precision eroded significantly when Apple's App Tracking Transparency framework restricted cross-app data collection. Meta lost a significant portion of its third-party signal.

In response, Meta pivoted its own playbook. The recommended approach is now broad targeting — let the algorithm find your audience using its remaining signal, your pixel data, and your creative performance as a targeting input. Meta's own guidance increasingly suggests letting the machine decide, rather than constraining it with narrow audience definitions.

In practice, this means the creative has become the real targeting mechanism. An ad that resonates with your ideal customer is the most reliable way to reach more people like them. The algorithm learns from engagement, conversion, and CAPI signals — and uses that to expand reach to similar users. The "targeting" isn't in the audience definition; it's in the creative itself.

Where they're different

Despite the structural similarities, there are meaningful differences that should inform how you use each channel.

The intent context

Demand Gen runs on Google's ecosystem — YouTube, Discover, Gmail. Even though users aren't searching, they're in a Google environment. Someone watching a YouTube review of your product category has implicit intent context that a Meta user scrolling their feed doesn't. This makes Demand Gen particularly effective for audiences who are already in a research or consideration mindset, even if they haven't yet converted it into a Search query.

The audience data pool

Google's audience data is fundamentally different from Meta's. Google knows what people search, what they watch, what they read in Gmail, and — through the Chrome ecosystem — much of their browsing behaviour. Meta's data is primarily behavioural within its own walled garden, enriched by whatever first-party and advertiser data it can collect. Neither is inherently better; they're different audience graphs that reach people differently.

Attribution and view-through

Meta's view-through attribution — a conversion counted when someone saw your ad but didn't click — is a perennial source of over-reporting. It's a legitimate attribution signal (exposure does drive action) but it's easily inflated to the point of claiming credit that isn't there. Demand Gen's attribution is more conservative by default, though Google's multi-channel attribution models have their own complexity.

The practical implication: don't compare cost-per-conversion numbers between Meta and Demand Gen without normalising for attribution window and conversion type. They're measuring different things.

Frequency and audience warm-up

Meta builds frequency faster within a defined audience cohort. If you're retargeting a specific list on Meta, the algorithm will find high-frequency exposure quickly — which is useful for time-sensitive campaigns but destructive for always-on retargeting without creative rotation. Demand Gen's frequency management is different; the audience pool is typically broader and the frequency build-up is slower. For retargeting, Meta is often more aggressive and requires more active frequency management.

What this means for your strategy

The practical implication of the Google/Meta convergence is that the skills, infrastructure, and thinking required to run these channels well are increasingly shared. If you're running Demand Gen well, the creative-first thinking you're applying there should inform how you approach Meta. If you've invested in CAPI for Meta, the same first-party data infrastructure that powers CAPI will also improve your Google audience signals and offline conversion imports.

The biggest risk is managing them in silos — separate agencies, separate reporting, separate creative thinking. When paid social and paid search operate independently, you end up paying twice for the same audience, learning at half the speed, and attributing the same conversion to both channels. The data and the strategy need to be connected.

This is also why "creative is the new targeting" applies to both platforms. The audience is whoever responds to your ad. On Search, that response is a click after a query. On Demand Gen and Meta, that response is engagement, view, or conversion after an interruption. The mechanisms are different, but the underlying principle — that great creative finds its own audience when given enough room and signal to work — holds for both.

"The audience is whoever responds to your ad. Your job isn't to define the audience precisely — it's to make something worth responding to."

If you're managing both channels and treating them as independent platforms with independent strategies, it's worth asking: are your creative learnings from Meta informing your Demand Gen asset strategy? Is your CAPI implementation also improving your Google offline conversion data? Are you attributing across both channels or counting the same customer twice?

These are the questions that the convergence between Google and Meta raises. They don't have simple answers, but asking them is how you get ahead of advertisers who are still treating 2026 like it's 2018.