Should we run YouTube and Demand Gen for direct sales or for awareness, and how do we judge them?
Our answer
Buy YouTube for direct sales through Demand Gen on a purchase goal. Keep reach and view campaigns for new reach or retail lift. Judge both by a holdout on total sales, not Google-reported ROAS.
- Confidence
- Medium
- Strongest proof
- 12 Google rules
- 12 datasets, 6 practitioners, 1 opinion or claim
- Sources
- 11 sources
- 31 statements
- G
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentDemand Gen returned more in sales tests than reach setups. It returned 0.98 times a brand's other channels against 0.63 times for reach and awareness setups (Haus). Measured found $1.81 incremental ROAS for conversion Demand Gen against $1.55 for reach or view.
Google's own reporting undercounts YouTube. Haus finds 2.3 times the site sales Google Ads credits, with 24% of the lift landing after the ads stop. Stella adds at least 2 weeks after the flight for YouTube and Demand Gen tests.
View-through conversions are a loose count. Google counts one from an impression with 1 pixel on screen, turns view-through bidding on for new Demand Gen campaigns, recommends a 1 day window and does not export these conversions to other systems.
Our judgment
- Buy YouTube for direct sales through Demand Gen on a purchase conversion goal, and bid it the way Choosing a Google bid strategy sets for other campaigns. Haus also puts blended setups at 0.68 times, and Brett Curry runs online store sales mostly through Demand Gen. Build conversions with value on Maximize conversions or Target CPA first, as Google says, and move to value bidding once Demand Gen's own threshold is met.
- Keep reach and view campaigns for jobs that judging on sales undersells, such as new reach or retail lift, and give them a longer test. Those jobs are reaching people who don't know you yet or lifting Amazon and retail sales. Haus's pure awareness tests ran about 6.9 weeks against 5.5 for conversion and put post-flight value at 2.2 times the brand's other channels, and Brett Curry leans to view and reach for retail lift and net-new reach at large scale.
- Judge YouTube and Demand Gen by a holdout on total sales, read at least 2 weeks after the flight against your breakeven rather than zero lift. Google's Conversion Lift covers Demand Gen, and Haus says one test is a read, not a verdict.
- Treat view-through conversions as reporting, not proof, and decide the setting before launch. Set the window to 1 day or turn view-through bidding off, and keep it on only where a lift test confirms the campaign. Measured, which favors it, pairs it with an incrementality test.
- Use blended reads of total sales against YouTube spend changes only as a weak signal where you're too small for a holdout, and hold any conversion factor loosely until your own test replaces it. Blended reads mean total sales, and Brand Search and Performance Max cost against YouTube spend changes, with every other change written down. Fospha's figures compare brands that chose their own spend, and AdwService found in one store that an apparent Demand Gen lift came from market growth, tighter bids and feed changes. The Q1 2026 benchmark from Common Thread Collective (CTC) works out to about 1.1 times (2.04 times incremental against 1.86 times reported, our arithmetic) where Haus's typical test finds 2.3 times.
Our judgment, built from Haus, Measured, Brett Curry, Google, Stella, Relevant Audience, Fospha, AdwService and Common Thread Collective: our reading of these sources, not a finding from any one of them.
What the sources say
31 statements from 11 sourcesMenachem Ani, an agency owner, says one client's 243 extra add-to-carts came almost entirely from a Demand Gen campaign on kids' YouTube channels. Add to cart is a soft signal the system fills cheaply.
“Add to cart is a soft signal, and the system will find the cheapest way to produce it.”
Google says view-through conversion bidding is on by default for new Demand Gen campaigns, counting views of 1 pixel on screen. It recommends a 1-day window and warns longer ones inflate reported conversions.
“The recommended window for primary VTCs is 1 day. If you set it at a higher window, you can expect a surge in primary VTCs reported.”
Google counts an engaged-view conversion after 10 seconds of a skippable in-stream ad or 5 seconds of an in-feed or Shorts ad, and includes them in the Conversions column.
“You’ll find Engaged-view conversions included in both the “Conversions” and the “All conversions” columns”
Google says Demand Gen value-based bidding needs 50 conversions with value in 35 days (10 in the last 7), or 100 across the account. It advises a tROAS 20% below your historical average and no changes for 14 days.
“Set tROAS 20% lower than your historical average ROAS to initiate your campaign.”
Google's Demand Gen A/B experiments test creative, audience, feed or bidding, not budget, with a 50% split. Results need at least 50 conversions per arm, so Google suggests shallow goals like Add to Cart.
“We don't recommend testing the budget as a variable at this time.”
Google says view-through conversions exclude people who interacted with your other ads and, outside a few campaign types, are for reporting only. A longer window usually records more of them.
“A longer view-through conversion window will usually increase the number of view-through conversions Google Ads records.”
Google advises Demand Gen budgets of at least 20 times your target CPA, and says at least $100 a day per ad group helps avoid a cold start. Ad groups with more than 50 conversions perform better.
“a minimum daily budget of at least $100 USD per day, per ad group will help avoid a cold start.”
Google's New Customer Value goal bids higher for new customers and is recommended for all purchase advertisers. New Customer Only suits strict acquisition budgets and needs a separate campaign for existing customers.
“You should set up a separate campaign to reach existing customers when using this mode.”
Haus's geo tests find YouTube Demand Gen returns about the same as a brand's other channels (0.98x), and a brand's first YouTube test now lands at 0.92x, up from 0.64x. Haus calls these comparisons directional.
“In the Demand Gen era, Demand Gen tests land right at parity with the brand's other channels (0.98x).”
Haus says its geo tests found YouTube drove about 2.3x the DTC sales Google Ads reports, and about 24% of lift came after the flight ended. Test and brand counts are not given.
“one test is a read, not a verdict.”
Haus says the typical brand's YouTube geo test finds 2.3x the DTC sales the ad platform credited (80% interval 1.7x to 2.9x), and about three in four tests land above the platform's number.
“None of these numbers are yours, and the only YouTube result that should move your budget is the one measured on your own data.”
In Haus YouTube geo tests, a typical 24% of lift arrived after the ads stopped, and counting that window raised typical ROI about 32%. A day after a flight, YouTube still booked about 64% of an in-flight day's lift.
“In the typical test, 24% of the lift landed in the post-treatment window.”
Haus says about half of YouTube tests are built to detect lifts under 4%, and brands that missed a benchmark on a first test beat it on the retest 44% of the time, so one test is not a verdict.
“one test is a read, not a verdict.”
Google's user-based Conversion Lift needs a Google rep, at least 1,000 observed conversions and a $5,000 minimum budget. Google allows 7-day studies but recommends more than 14 days.
“We've found up to a 17% drop in Absolute Lift in studies with a long conversion lag that run for less than 14 days.”
Google's data-driven attribution, the default for most conversion actions, splits credit across your Google ad interactions. Any volume qualifies, but Google recommends 200 conversions and 2,000 interactions in 30 days.
“Depending on data availability, the last click and data-driven attribution models can have the same results in certain situations.”
Brett Curry, whose agency runs YouTube ads, prefers fast-paced direct response ads over 30 seconds and says YouTube looks terrible on one-day click ROAS, so judge it another way.
“It's going to be terrible. If that's all you measure, then we should not do that. Do something else”
Relevant Audience, relaying Google, says Demand Gen view-through bidding becomes video-only, expands to the Display Network, and is on by default for new campaigns. Google gave no rollout dates.
“Separate the conversion column you report on from the one you optimise on.”
Fospha's 2025 analysis of 127 retail brands found those scaling YouTube and Demand Gen together saw 32% stronger blended revenue growth year over year. It compares brands' own choices, not a test.
“The 32% / 145% growth figures are for brands scaling both together, so the data doesn't isolate YouTube's standalone contribution.”
Google Analytics now lets you set click-through conversion windows to any whole number from 1 to 90 days and engaged-view windows from 1 to 30 days (previously fixed at 3 days).
“Edit EVC conversion windows to any integer from 1 to 30 days (previously fixed at 3 days).”
Fospha, a measurement vendor, found across 25 retail ecommerce brands that those putting 10 to 20% of Google budget into Demand Gen had 100% higher Google ROAS than those at 0 to 5%. That is association, not cause.
“the strongest results come from funding it with incremental investment rather than reallocating from PMax or Search.”
AdwService, which sells Google Ads management, found no significant Demand Gen effect on sales in one e-commerce account over about 18 months. It says only a holdout can prove incremental value.
“we found no statistically significant effect of Demand Gen on conversion rate, sales, or lower-campaign attribution value.”
Brett Curry, relaying Haus's study of 190 tests, says in-platform reporting undercounts YouTube by about 70%. One retail matched-market test showed about 25% sales lift over eight weeks.
“If you see a one ROAS in platform for YouTube, you're probably actually getting a 3.4 ROAS”
In a real client's YouTube geo test design (about $12.4K a month, 17 weeks), 4 control cities could detect a 21.5% lift and 10 cities 15.0%, so a true 10% lift would be missed.
“If your team would reallocate budget at a +10% lift but not at a +5% lift, the test needs to distinguish those two outcomes.”
Stella, which sells geo testing, found median incremental ROAS across 225 tests of online consumer brands of 2.98x for Performance Max, 2.92x Meta, 2.17x YouTube, 1.86x Shopping and 0.70x branded search.
“For planning purposes, we recommend discounting these benchmarks by 15-20% to establish conservative targets for new testing programs.”
Across 225 geo tests on its platform, Stella, which sells testing, found a median 2.31x incremental ROAS. It suggests 4 to 6 weeks for most brands and skipping tests that can't get a few hundred conversions.
“The confidence interval matters more than the point estimate”
Measured, which sells incrementality testing, found 86% of 68 Demand Gen studies showed significant sales lift; for ecommerce, conversion-optimized Demand Gen returned a median $1.81 incremental ROAS.
“In-platform reporting can underreport the incremental value of YouTube placements by as much as four times.”
Measured, which sells incrementality tests, says across hundreds of tests 62% of view-optimized YouTube conversions were net new against 46% for Demand Gen, and urges turning on view-through optimization.
“Pair it with an incrementality test to validate the new signal is driving results.”
Google announced Demand Gen Uplift Experiments, automated A/B tests of adding Demand Gen to other campaign types. It cites a Google-commissioned Nielsen study: on average 10% higher ROAS adding it to Search and PMax.
“An automated A/B testing infrastructure”
Google announced at Marketing Live 2026 that Demand Gen campaigns can optimize toward view-through conversions alongside click-based ones. It gives no figures; the benefit claims are Google describing its own product.
“A new bidding enhancement for Demand Gen that allows advertisers to include and optimize for View-Through Conversions (VTC) alongside existing click abased conversions.”
CTC's Q1 2026 benchmark of 299 brands, using its incrementality factors, puts Google brand search at 19.1x platform ROAS but 5.7x incremental, and Meta acquisition at 1.83x reported versus 2.07x.
“Google Brand search reports 19x ROAS but delivers just 5.7x after incrementality adjustment (0.30x IF).”
Common Thread Collective's Q1 2026 benchmark shows YouTube at 0.90% of spend with a 1.10x incrementality factor and 2.04x iROAS. Its text gives different figures, calling YouTube its most underinvested channel.
“the most underinvested channel at just 0.36% of total spend despite carrying a 1.3x incrementality factor”
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer