According to Measured, 72% of incremental conversions on YouTube come from new customers. That's a striking number, and Google is leaning hard into it with June's Demand Gen feature drop. But before anyone restructures a media plan around a single stat, it's worth pulling apart what's actually shipping, what it changes for B2B operators, and where the trade-offs hide.
What's in the June drop
Three updates, none of them massive on their own, but together they signal where Google wants Demand Gen to go.
1. Expanded video resizing. Demand Gen will support broader aspect ratio transformations: vertical to square, vertical to landscape, square to landscape. The practical upside is obvious. If your creative team produces one vertical hero video, the system can now reformat it across more placements (YouTube, Shorts, Discover, Gmail, GDN) without manual re-edits. The risk? Automated resizing can crop key elements. If your video relies on text overlays or tight framing, you'll need to QA the outputs before trusting the machine.
2. Gemini-powered creative insights. When you're selecting image and video assets for a campaign, Gemini will surface automated recommendations on how to optimize creative for YouTube. Think of it as a pre-flight checklist generated by AI. Useful for catching obvious misses (wrong aspect ratio, low-contrast thumbnails). Less useful for the harder creative questions: does this hook land in the first three seconds? Does the value prop match the audience segment? Those still require a human with context.
3. Web-to-App Acquisition Measurement. You can now see how Demand Gen campaigns drive new app installs, giving a fuller picture of performance. For B2B SaaS teams with a meaningful app install motion, this closes a measurement gap. For everyone else, it's a nice-to-have.
Where this fits in a B2B SaaS media plan
The context matters more than the features. Paid acquisition's share of B2B SaaS pipeline has dropped from roughly 34% to about 26% over recent cycles, while owned channels (organic, content, AEO) climbed from 22% to 27%. Meanwhile, CPL inflation hasn't slowed: LinkedIn ads are up around 24%, Google up approximately 19%. Top-quartile SaaS teams now attribute 41% of qualified pipeline to owned channels combined.
So the question isn't "should we try Demand Gen?" The question is: can Demand Gen help maintain paid pipeline contribution without scaling cost proportionally? And the honest answer is: maybe, if you do the pre-work.
That pre-work has three parts. First, your first-party data needs to be clean enough to build useful lookalike segments. Experts consistently recommend lookalike audiences based on past purchasers, site visitors, or YouTube viewers to expand reach while maintaining qualification. Second, your creative mix should include both video and image assets (campaigns using both typically outperform single-format setups). Third, and this is the one teams skip: your landing page experience has to be strong. Traffic without a post-click experience that converts is just expensive awareness.
The measurement problem nobody's solving with features
None of June's updates address the core measurement challenge for B2B Demand Gen: connecting ad engagement to qualified pipeline. Google's Offline Conversion Tracking integration exists for this purpose, allowing you to import lead qualification data from your CRM back into the campaign. But it only works if your CRM hygiene is solid and your RevOps team has built the feedback loop. If you can't reliably tag which leads became sales-accepted opportunities, you're optimizing toward the wrong signal.
Google is also shifting some Demand Gen campaigns optimized for view-through conversions from CPC to CPM billing. That changes your cost dynamics and reporting expectations. If your team is forecasting based on CPC benchmarks, recalibrate now or you'll misread performance for the first few weeks.
The broader point: platform dashboards can't prove incrementality on their own. Lift studies (brand lift, search lift, conversion lift) are the only way to validate whether spend is creating demand or just capturing it. Treating last-click as proof of anything is a recurring failure mode.
What to actually do this week
Setup: If you're already running Demand Gen, audit your creative library. Do you have both video and image assets loaded? Are your videos formatted for multiple aspect ratios, or are you relying on auto-resize? QA the resized outputs once the feature rolls out.
Hypothesis: If we add Gemini's creative recommendations as a pre-launch checklist (not an override of creative strategy), then asset rejection rate will drop and time-to-launch will shorten, because we'll catch formatting issues earlier.
Success metrics: Primary: cost per qualified lead (not cost per lead). Secondary: creative approval cycle time, landing page conversion rate. Guardrail: if CPA rises more than 20% during the first two weeks of a new creative mix, pause and diagnose before scaling.
Stop-loss: If after 14 days and sufficient spend (2x your target CPA × 50 conversions), qualified pipeline contribution hasn't improved directionally, reallocate budget to the owned-channel experiments that are working.
Video is projected to hit 82% of internet traffic by 2026, and the standard recommendation is to allocate 60% of video budget to distribution rather than production. Demand Gen is a distribution lever. But a lever only works when there's something worth distributing on the other end.
The June drop gives you better formatting tools, an AI creative assistant, and a new measurement surface. None of those fix bad inputs. The teams that win with Demand Gen this quarter will be the ones who already have clean data, strong creative, and a post-click experience that earns the conversion. The features just make it slightly easier to scale what's already working.
