YouTube Demand Gen Campaigns in 2026: The Underused Acquisition Channel — Shorts, Lookalikes, the March Lookalike Update, and the 9-Move Operator's Playbook

A D2C founder told me last quarter that "YouTube Ads doesn't really work for us." Spend: $14,000/month. Conversions: 38. CPA: $368. Same products were running profitably on Meta at a $89 CPA. The conclusion looked obvious — YouTube was the wrong channel.

We pulled the account. Single Demand Gen campaign. One 30-second video, horizontal 16:9 only, no vertical assets for Shorts. Two static images. Four headlines, two descriptions. Ad Strength rated "Average." Targeting set to a single broad audience with no Lookalike segment. Bidding strategy frozen on Maximize Conversions for six months despite having more than enough data to switch to Target CPA. Force-migrated from a Video Action Campaign in July 2025 with no post-migration audit. We didn't change the channel. We rebuilt the campaign over four weeks. Asset bundle expanded to Excellent Ad Strength: 9 images across all three aspect ratios, 4 vertical 9:16 videos for Shorts, 5 headlines and 5 descriptions, business name and clear CTA. Two Lookalike segments built from the top 500 customer-fit profiles, configured for AI-powered signal mode after the March 15 update. Switched bidding to Target CPA at $120 (a deliberate stretch above current performance). Pivoted creative mix to UGC-style Shorts and product demonstration videos. CPA dropped from $368 to $94 within 30 days. Volume tripled at the same monthly spend. The channel works fine. The playbook just needed to be from 2026, not 2024.

This is the most common pattern we see in YouTube Ads accounts in 2026. Google has rebuilt this product line twice in 24 months — first replacing Discovery campaigns with Demand Gen in late 2023, then force-migrating every Video Action Campaign to Demand Gen in July 2025, then shipping the March 15, 2026 Lookalike segment update that just changed how prospecting works. Most agencies are operating on playbooks that predate at least one of those shifts. The result: a generation of accounts blaming "YouTube doesn't work" when the platform has been quietly outperforming and the operating model is simply outdated.

This piece is the operator's playbook we use to run YouTube Demand Gen campaigns in 2026. It covers the platform's actual current state, the 9 levers that separate scaling accounts from plateaued ones, the post-March Lookalike configuration, the Excellent Ad Strength asset bundle, the Shorts-first creative spec, and the 60-day rollout for migrating existing accounts.

The state of YouTube Ads and Demand Gen in 2026

YouTube is no longer a brand-only play. The numbers reframe the channel's role in any modern paid acquisition stack:

Reach: Demand Gen reaches over 3 billion monthly active users across YouTube, YouTube Shorts, Discover, and Gmail. YouTube Shorts hit 200 billion daily views in Q2 2025 (Google's official Q2 figure was 50B daily on Shorts alone — both numbers depend on counting methodology). YouTube is now the most-watched streaming platform on connected TVs in the US, surpassing Netflix per Nielsen December 2025 data. More Americans watched TV on YouTube than any other platform that month.

Performance vs predecessor: Nielsen analysis found Demand Gen delivers 58% higher ROAS than Video Action Campaigns (the predecessor format Google force-migrated in July 2025). LG Electronics reported 24% higher conversion rates running Demand Gen vs equivalent Meta campaigns. Demand Gen advertisers in 2025 saw a 26% year-over-year increase in conversions per dollar spent.

Best-practices payoff: Google's internal analysis of Demand Gen advertisers found that those using at least 3 of the 4 recommended best practices saw 40%+ more conversions on average. Campaigns running both video and image assets together generated 20% more conversions at equivalent CPA than video-only campaigns.

Real case study: Cropp, the Gen-Z apparel brand owned by LPP Group, achieved a 50%+ ROAS increase and 4.5% conversion lift among women shoppers through a properly configured Demand Gen campaign — diverse creative across formats, optimized targeting plus Lookalike segments built from existing customer data, and conversion-focused bidding strategies with appropriately set targets.

Trustworthiness: Google-commissioned Ipsos research found viewers rated YouTube ads 16% more trustworthy, credible, and honest than ads on competing video and social platforms. The trust delta is real and structural — it's why YouTube Shorts ads convert disproportionately better than equivalent TikTok content for considered-purchase categories. The channel is mature, well-instrumented, and underserved by most performance marketing operators. CPMs sit between TikTok ($3-5) and Meta ($10-15), with conversion quality and trust signal that exceeds both for the right categories.

What changed in 2025 and 2026

Three structural shifts define the platform now. Most agency playbooks predate at least one.

Video Action Campaigns force-migrated to Demand Gen in July 2025. Google deleted the entire Video Action Campaign product line and migrated every active campaign to Demand Gen automatically. Most accounts were paused after migration for advertiser review; many remain paused or running on default settings nine months later. The structural change matters: Demand Gen handles creative differently, optimizes across more surfaces, and has different reporting. A Video Action Campaign that was working before migration is not the same campaign after migration, and treating it as if it is produces the kind of underperformance pattern we showed in the opening.

The March 15, 2026 Lookalike segment update. Lookalike segments in Demand Gen transitioned from strict targeting (where only matched lookalike users see ads) to AI-powered audience signals (where lookalike data informs the algorithm but does not strictly gate delivery). This is the most significant structural change in Demand Gen prospecting since the campaign type launched. Advertisers who don't update their audience configuration after this transition see their old strict-targeting setups behave differently than expected — typically with broader reach than intended and audience signal quality that's effectively diluted unless the configuration is rebuilt. The piece is being published one week after this update went live; most accounts haven't audited.

Excellent Ad Strength became the de facto requirement. Google's February 2026 best practices guide explicitly correlates Excellent Ad Strength with substantially higher campaign performance — the Cropp case study and the broader 40%+ conversion lift data both depend on hitting it. Excellent Ad Strength requires a specific asset bundle that most accounts don't ship: at least 3 unique images per aspect ratio (landscape 1200×628, square 1200×1200, portrait 960×1200), at least 1 video, vertical 9:16 video for Shorts placements, 5 headlines (40 chars max), 5 descriptions (90 chars max), business name, and clear call-to-action. Accounts shipping minimum-viable assets cap their performance ceiling regardless of how well the campaign is otherwise configured.

The platform also expanded Smart+ features to Demand Gen (channel controls now let you opt out of GDN if it's not performing), added carousel and video-plus-image combinations, and improved cross-channel reporting between YouTube, Shorts, Discover, and Gmail surfaces.

The 9-Move Operator's Playbook

This is the framework we use across client accounts. Sequenced from foundational to advanced.

1. Achieve Excellent Ad Strength on every active campaign. This is the highest-leverage move and the one most accounts skip. The required bundle: 3+ unique images per aspect ratio (landscape, square, portrait), at least 4 videos with at least 1 vertical 9:16 for Shorts, 5 headlines (40 chars), 5 descriptions (90 chars), business name, and clear CTA. Accounts shipping minimum assets sit at "Average" or "Good" Ad Strength and miss the 40%+ conversion lift Google's data shows. The asset volume requirement is similar to what Meta's Advantage+ and Google's PMax now demand — creative production volume is the binding constraint across every major paid platform in 2026.

2. Configure Lookalike segments correctly after the March 15 update. Three reach levels: Narrow (top 2.5%), Balanced (top 5%), Broad (top 10%). After the March 15 transition to AI-powered signal mode, the configuration logic shifted: the Lookalike segment now informs the algorithm rather than gating delivery. Best practice in 2026: build separate Lookalike segments from each high-value seed (top 500 customers by LTV, repeat purchasers, recent converters) and run them as audience signals alongside Custom Intent and In-Market segments rather than as the sole targeting layer. Narrow Lookalikes (2.5%) for high-value customer prospecting; Broad (10%) for top-of-funnel reach where volume matters more than precision.

3. Use the bidding ladder, not a single bid strategy. Demand Gen rewards a sequenced bidding approach: Maximize Clicks for the first 14 days to gather data quickly. Switch to Target CPA once you have ~30 conversions in a 30-day window — this gives the algorithm enough data to optimize against a target. Move to Target ROAS once stable, especially for ecommerce brands with a wide variety of price points. Pair the chosen strategy with Standard pacing to prevent burning budget early in the day. Most underperforming accounts we audit are stuck on Maximize Conversions for months despite having data to switch — the bidding strategy is leaving meaningful efficiency on the table.

4. Default to vertical-first creative for Shorts placements. YouTube Shorts is now the highest-growth placement in Demand Gen, and you can't manually select "Shorts only" — Google distributes assets based on creative format and performance. If you only upload horizontal 16:9 videos, you miss Shorts inventory entirely. The 2026 spec: 9:16 vertical mandatory at 1080×1920, square 1:1 at 1080×1080 secondary, horizontal 16:9 at 1920×1080 for in-stream and Watch Next placements. Hook in the first 3 seconds to stop the scroll. Keep videos 15–30 seconds — anything longer typically gets relegated to YouTube Home feed where engagement drops.

5. Combine video and image assets in the same campaign. Google's data is decisive here: campaigns running both video and image together generate 20% more conversions at equivalent CPA than video-only campaigns. The placement logic differs by surface — Discover Feed and Gmail favor static images, YouTube Shorts and in-stream favor vertical video, YouTube Home feed favors square or vertical static plus video carousel. Shipping both formats lets the algorithm match the right asset to the right placement. Accounts running video-only campaigns are leaving the static-asset placements entirely uncovered.

6. Use channel controls deliberately. Demand Gen now lets you select which channels to include — YouTube, Discover, Gmail, Google Display Network. Default behavior includes everything. The strategic move: launch with all channels enabled to gather data, then exclude underperforming channels after 4-6 weeks of learning. GDN often performs the worst on quality of conversion (cheap clicks, low intent); Gmail typically performs best for retargeting. Discover Feed sits in the middle. Channel-level performance varies dramatically by category; let the data, not the default, decide your placements after the initial learning period.

7. Build proper Custom Intent audiences from your Search keywords. Custom Intent audiences let you target people who've recently searched for specific terms on Google. The highest-leverage move: pull your top-converting keywords from Search and Shopping campaigns, then build Custom Intent audiences from those exact queries. You're handing Demand Gen the exact intent signals your manual Search campaigns already proved work. This produces audiences that are functionally "people about to buy from a brand like yours" rather than generic in-market segments. Most accounts skip this step and use stock In-Market segments instead — you're leaving precision on the table.

8. Run 4-6 week minimum decision windows. Demand Gen's algorithm has a longer learning phase than Search or Shopping because the conversion signals are sparser and the placements are more diverse. Google itself recommends 2-4 weeks of data before optimization decisions. The minimum reliable window is 4 weeks; 6 is more reliable. The minimum monthly budget that produces meaningful learning is roughly $3,000-5,000 per campaign — below that, you don't get enough conversions per week for the algorithm to learn properly, and you'll draw false conclusions from noisy data.

9. Treat Demand Gen as the mid-funnel partner to PMax, not a substitute. PMax captures existing demand on Search and Shopping placements. Demand Gen creates demand on visual feed surfaces. They work different parts of the funnel and the right operating model runs both. The accounts winning in 2026 use PMax for bottom-funnel capture (high-intent search queries, shopping comparisons) and Demand Gen for top- and mid-funnel demand creation (visual feeds, Shorts, Gmail, Discover). The two campaign types feed each other: Demand Gen creates demand that PMax later captures. Running them as competing alternatives produces sub-optimal allocation. Running them as complementary layers compounds.

Demand Gen vs Performance Max: when to use which

A common question we get from CMOs: "We're already running PMax. Do we need Demand Gen too?"

The honest answer is yes, for most accounts above $25K/month in Google Ads spend. They serve different jobs.

Use PMax when: you're capturing existing search and shopping intent, you have limited creative production capacity (PMax generates AI variations from your assets), you want maximum automation across Google's surfaces, your products are familiar enough that branded search drives meaningful demand. PMax is dominant on Search and Shopping placements where users are actively comparing.

Use Demand Gen when: you're trying to create demand at the top and middle of the funnel, your products are visually compelling and benefit from feed-based discovery, you have high-quality video and image assets to feed the algorithm, you want more granular control over audience targeting and channel placements than PMax allows. Demand Gen is dominant on visual feed surfaces (Shorts, Discover, Gmail, YouTube Home).

The right operating model is both. Allocate 60-70% of Google budget to PMax for bottom-funnel capture and Search-adjacent demand. Allocate 20-30% to Demand Gen for top-funnel demand creation and visual discovery. Allocate 10-15% to standard YouTube and Display campaigns where granular control matters. The mix shifts by category — visual-first products (apparel, beauty, food, lifestyle) push toward 30-40% Demand Gen; commodity products with high purchase intent push toward 70-80% PMax.

For the deeper PMax operating playbook, see Performance Max in 2026 — the diagnostic framework and 10 levers carry over directly.

The 60-day rollout for an existing Google Ads account

The sequence we use to migrate accounts from underperforming Demand Gen to a structured program.

Days 1-15: Foundation. Audit current state — Ad Strength rating on every campaign, Lookalike configuration, bidding strategy, asset inventory, channel performance breakdown. Audit force-migrated VAC campaigns specifically — many remain paused or running on default settings. Build out the Excellent Ad Strength asset bundle: 9+ images across all three aspect ratios, 4+ videos including vertical 9:16, 5 headlines, 5 descriptions. Wire up Conversions API and Enhanced Conversions if not already running (covered in the SST field guide).

Days 16-30: Live tests. Launch 2-3 Demand Gen campaigns: one prospecting campaign with broad Lookalike audience signal (top 5%) and Custom Intent from converted Search keywords, one retargeting campaign on website visitors and engaged users, one new-customer-focused campaign using exclusion lists for existing customers. Run on Maximize Clicks for the first 14 days to gather data, then switch to Target CPA at 110-120% of current platform CPA.

Days 31-60: Optimization and scale. Once each campaign has 30+ conversions in a 30-day window, switch to Target CPA. After another 14 days of stable performance, evaluate Target ROAS for revenue-driven categories. Audit channel performance — exclude GDN if quality is poor, double down on Shorts if engagement is strong. Refresh creative assets every 4-6 weeks to combat fatigue. By day 60, the program should be producing CPAs at or below Meta and Google PMax, with substantially better top-of-funnel demand creation than either platform delivers in isolation.

By day 90, Demand Gen typically becomes the second-largest channel in a well-run Google Ads account behind PMax — outperforming standard Display and brand-only YouTube spend by substantial margins.

What this means for your account this quarter

YouTube Demand Gen in 2026 is the most underused acquisition channel in performance marketing. The platform has matured. The benchmarks have been published. The case studies show 50%+ ROAS lifts and 40%+ conversion improvements for accounts that operationalize correctly. Google has invested heavily in making the product work — and Most agencies haven't updated their playbook to match the investment.

The brands winning are not the brands spending the most on YouTube. They're the brands that hit Excellent Ad Strength, configured Lookalike audiences correctly after the March update, sequenced their bidding ladder, defaulted to vertical Shorts-first creative, ran 4-6 week decision windows, and treated Demand Gen as a complement to PMax rather than a substitute. The 9 moves on the playbook above are not advanced — they're foundational. Most accounts haven't shipped them. If you'd rather have someone audit your YouTube account, restructure for Demand Gen, build the Excellent Ad Strength asset bundle, configure post-March Lookalike segments, and ship the 60-day rollout — that's part of the work we do at Praxxii Global. Across our Google Ads accounts in 2026, the average lift from a structured Demand Gen migration has been 52% on ROAS and 38% on CPA reduction within 60 days, against unchanged spend. That isn't a creative breakthrough or a clever audience hack. It's the disciplined operational work most accounts haven't gotten to yet — and the platform best practices Google has been publishing for the past 18 months that most accounts haven't implemented.

YouTube reaches more US households on connected TV than Netflix. Demand Gen reaches 3 billion monthly active users across the most attention-rich feeds in Google's ecosystem. The trust score is the highest in paid social. The conversion data is the strongest in mid-funnel. The channel works. Pick whichever moves on the playbook above will close the biggest gap in your current setup. Ship them this week. The compounding starts there.