Demand Gen on Google Ads in 2026: The Real Verdict After Two Years

When Google retired Discovery Ads in late 2023 and replaced it with Demand Gen, the pitch sounded like a marketer’s dream: native video formats, lookalike targeting, built in A/B testing and access to the full YouTube inventory: on paper, the promise was gorgeous. Two years and a lot of media budget later, we can finally give an honest verdict, from what our own client accounts showed us.

Discovery Ads Is Dead, Long Live Demand Gen

The forced migration annoyed plenty of advertisers, and that reaction was fair. Nobody enjoys watching a channel that performs get unplugged overnight.

The first shift is about formats: where Discovery stuck to fairly tame image units, Demand Gen campaigns open the door to video and, above all, to YouTube Shorts. You reach your audience where it spends time, inside a vertical feed people scroll dozens of times a day. The second gain is targeting: similar segments, or lookalikes, let you chase profiles close to your best customers and widen an already qualified audience. Add native A/B testing, which finally allows experiments without manual workarounds, and you get a tool that is clearly more mature than its predecessor.

During 2025, Google also rolled Google Display inventory into Demand Gen, stretching reach across a huge slice of the open web. In the same wave, the default bidding strategy moved from target CPA to target ROAS, a clear signal that Google wants Demand Gen to line up against Meta and TikTok on performance, not just awareness.

The real issue was never the tool, it was the way many advertisers launched it, exactly like Discovery, with the same static creatives, the same delivery logic and the same expectations. The result was disappointing numbers and a slightly unfair trial for a format that keeps its promises once it is fed properly.

How Demand Gen Performs in 2026, Vertical by Vertical

Let us get concrete, because this is where market averages become misleading. Demand Gen campaigns do not deliver the same results across every industry, and our field observations show sharp gaps.

In fashion and lifestyle ecommerce, this is clearly the ideal playground. Shorts crush it, product video works very well, and the format fits impulse driven buying beautifully. Travel follows close behind, with a very strong top of funnel, since nothing beats a (jet2 holiday, sorry I had to do this one) good video sequence to spark the urge for a destination. Finance and insurance do a decent job, as long as you accept creatives boxed in by regulation, which caps some of the creative upside. Real estate stays functional but calls for close ROI monitoring, because a long decision cycle muddies attribution.

Then comes the awkward case: B2B and SaaS. Let us be blunt, Demand Gen is simply not the right tool here. Reaching a CTO or a head of procurement in the middle of an entertainment feed is an uphill battle, and cost per lead climbs fast. For these sectors, concentrate budget on intent channels like Search. Discover our paid acquisition services focused on capturing demand that already exists.

The Golden Rule: No Video, No Point

If there is one lesson worth keeping from these two years, this is it. On Demand Gen, purely static formats badly underperform. The algorithm feeds on video, the flagship inventory is YouTube, and a plain fixed image simply drowns without ever catching an eye.

The wider market only reinforces this. Audiences have moved to video, and YouTube now reaches more than a third of the planet every single month. In other words, attention has relocated, and video is where you capture it now.

The takeaway is clear for any leader steering a budget. If you cannot produce video, even simple video, put that money elsewhere. A solid vertical Reel, a few format variations and a sharp hook in the first five seconds beat ten polished but frozen images every time.

You can use ou Creative Studio to create videos, built to make paid social convert.

Demand Gen or Meta Ads: The Wrong Question

We hear it constantly: should you choose between Demand Gen and Meta Ads? It is a false choice, and it makes you miss the point.

Meta keeps strengths that nobody takes away: interest based targeting stays remarkably fine grained, and the platform excels at converting directly, at the bottom of the funnel. Demand Gen, for its part, holds a trump card Meta can never replicate: the reach of YouTube. Billions of hours of video watched every day, an audience larger than a third of the world, that is exposure Facebook and Instagram simply do not cover.

So the winning strategy in 2026 is to orchestrate both rather than pit them against each other. Demand Gen owns the top of funnel on YouTube and Discover, builds the brand and creates the want. Meta then takes over with precise product retargeting and closes the sale. One generates demand, the other converts it.

The Mistakes We Still See Too Often

The first mistake we often see is mixing a traffic goal and a conversion goal inside the same campaign. The algorithm needs a single direction to optimise properly, so pick your priority and hold the line. The second concerns attribution: trusting view based conversions on a seven day window opens the door to a massive overstatement of results, you should stay on a short window of one day until you have proof to justify anything longer. The third mistake is launching lookalike audiences far too broad from the start, begin with around two to five percent similarity, then widen gradually.

Our Verdict After Two Years of Demand Gen

Demand Gen has found its place, and that already counts for a lot. It is not the big changing moment that was announced, but it is a solid, durable channel for advertisers who have the creative power to feed it.

The move to target ROAS and the added Display inventory show that Google keeps investing seriously in the format, which is reassuring for the long run. Still, success comes down to a simple equation: quality video, a clear goal and smart orchestration with your other levers. If video is still missing, produce it before you open the budget. At Donutz Digital, we have learned to treat Demand Gen for what it is, an excellent demand generator, as long as you do not ask it to do the job of an intent channel.

FAQ

Does Demand Gen really replace Discovery Ads in 2026?

Yes, the transition is complete. Google migrated every Discovery campaign into Demand Gen, then folded Video Action campaigns into the format during 2025. Discovery Ads no longer exists as such, and Demand Gen is now the single entry point for demand generation across YouTube, Discover and Gmail. You can review the current Demand Gen campaign guidelines straight from Google.

Can you run Demand Gen without a video budget?

Technically yes, since the format accepts image and carousel assets. In practice we strongly advise against it. Demand Gen campaigns without video clearly underperform, because the main inventory and the algorithm are built for video. Producing even one simple Short beats running static only.

What budget should you plan to test Demand Gen properly?

There is no magic number, but plan enough to fund both creative production and a real learning phase, usually a few weeks. The classic error is opening a small budget, cutting after ten days, then concluding the channel does not work when it never had time to exit its learning phase.

Does Demand Gen work for B2B and SaaS?

This is the sector where the tool disappoints most. Reaching a B2B decision maker in the middle of an entertainment feed stays difficult, and cost per lead climbs quickly. For these activities we prioritise intent channels like Search, keeping Demand Gen for very targeted awareness tests at most.

How do you measure a Demand Gen campaign correctly?

Focus on a short attribution window, ideally one day after the impression, to avoid overstating real impact. Then cross the data with your other channels, because Demand Gen often assists conversions that close elsewhere, on Search or direct. A multichannel read stays the only way to judge its true contribution.