What Google changed
Google's Demand Gen budget recommendation now sits at 10x your target CPA. The previous guidance was 20x your tCPA, which also carried a floor of $100/day.
The math is direct. If your target CPA is $20, the recommended minimum just dropped from $400/day to $200/day. At $15 tCPA, you're looking at $150/day instead of $300. For brands that have been watching Demand Gen from the sidelines because the cost to test was too high, this is the first meaningful change to that calculation in years.
There was no official blog post. The guidance updated in Google's help documentation. Two Octobers flagged it as part of their September 2026 platform roundup.
Why Google did it
Demand Gen has a participation problem. Google built the format as its YouTube-first answer to Meta's top-of-funnel reach — YouTube, Discover, Gmail, and Display inventory in one campaign type. But adoption among DTC and mid-market e-commerce accounts has stayed low relative to Search and PMax.
The most common reason media buyers cite is the budget threshold. At 20x tCPA, a brand with a $30 target CPA needed $600/day to run the format correctly. For accounts spending $3,000–$8,000/month total on paid media, that single campaign would have consumed a third to half of everything.
Google lowered the floor. The charitable read is that they have data showing 10x is enough for the algorithm to learn. The less charitable read: they need more advertisers participating to build a deeper auction. Both things can be true simultaneously.
What the lower floor doesn't fix
Budget was one barrier. It was not the only one.
Demand Gen still requires YouTube-native creative. A 9:16 video built for TikTok or Instagram Stories rarely transfers to the YouTube inventory that Demand Gen prioritizes. If you don't have 6-second bumpers, 15-second skippable ads, or 30-second assets built for connected TV, the budget reduction saves you money you couldn't spend effectively anyway. The production gap doesn't close automatically.
The view-through conversion problem also hasn't moved. Google's Demand Gen reporting includes one-day VTC — conversions attributed to ad exposure without a click. This inflates in-platform ROAS figures. When brands run incrementality tests or compare third-party attribution against Demand Gen's reported numbers, the gap is frequently significant. The budget floor changed; the measurement opacity didn't.
Then there's the learning phase. Even at 10x tCPA, you're looking at four to six weeks before the algorithm settles and you can make meaningful optimization decisions. For a brand running Demand Gen at $200/day, that's $5,600–$8,400 in spend before you have data worth acting on. The lower entry point reduces the financial risk. The time cost is identical.
The accounts this actually opens up
The 20x requirement effectively made Demand Gen a mid-market and enterprise format. Most accounts under $15,000/month in total ad spend couldn't justify a test at $400–$600/day without gutting the campaigns that were already working.
At 10x, that changes. A brand running Google Ads at $5,000–$10,000/month with a $20 tCPA can now run a Demand Gen pilot at $200/day without it dominating the rest of the account structure. That's a meaningfully different category of advertiser than the format previously served.
If you're in that range — established enough to have solid conversion data on Search or Shopping, but not yet at scale where Demand Gen is an obvious addition — this is the first time the format fits your budget without forcing a hard trade-off somewhere else.
Before you start, two checks
Use your actual CPA from the last 30 days, not your target CPA. If your account-level CPA is running at $35 but your tCPA is set at $20, your realistic minimum budget is $350/day, not $200. The recommendation is based on the target you set, not the performance you're getting. Use real numbers before committing budget.
Audit your creative library. If everything you have was built for Meta or TikTok, you don't have what Demand Gen needs. Running the format with square images and repurposed Reels is technically possible. It's also a reliable way to burn through the learning phase budget on impressions that won't convert efficiently at the format's ceiling. Native YouTube assets first, then turn on the campaign.
If you pass both checks, a four-week Demand Gen pilot against your current Meta awareness spend is a reasonable test to run before Q4 CPM pressure arrives. Demand Gen inventory is cheaper when fewer advertisers compete for it. That window narrows as adoption picks up.
If you want to see where your Google Ads budget is actually going before adding a new campaign type, Gromerce's free account audit shows the full picture in three minutes: gromerce.com/audit
The budget floor dropped. The creative bar and the attribution fog are right where you left them.
Sources: Google Ads Help Documentation, ALM Corp, September 2026
