Demand Gen

This page is updated every two months with current best practices for Google Ads Demand Gen campaigns. Demand Gen brings social-style visual advertising to Google inventory across YouTube, Discover and Gmail, and rewards strong creative and well-built audience signals. Each update draws on our own experience plus authoritative industry sources and verified real-time research. Bookmark this page and check back for the latest Demand Gen best practices. Each update includes worked examples with the arithmetic shown.

Last updated: September 15, 2026

In This Guide

  1. Executive Summary
  2. Benchmarks & Numbers at a Glance
  3. What Demand Gen Is
  4. Campaign Setup & Goals
  5. Creative & Ad Formats
  6. Audiences & Lookalikes
  7. Bidding Strategies
  8. Placements
  9. Measurement & View-Through
  10. Demand Gen vs PMax vs Video
  11. Common Mistakes to Avoid
  12. What Changed Recently
  13. References

1. Executive Summary

Google Ads Demand Gen campaigns are the primary vehicle for visually-driven, AI-optimized advertising across YouTube, Discover, and Gmail as of September 2026. Five principles separate high-performing accounts from mediocre ones:

  • First-party data is the foundation. Customer Match lists, website visitor segments, and offline conversion imports provide the signal quality that makes every downstream feature—lookalikes, optimized targeting, value bidding—work reliably. Weak data in means weak results out.[1]
  • Creative breadth is a system, not a single ad. Google’s AI needs a diverse pool of images and videos across vertical (9:16), square (1:1), and landscape (16:9) ratios—at least 3 unique assets per ratio—to match the right creative to the right placement and user. A single hero asset is not a strategy.[2]
  • Bid strategy must follow conversion data maturity. Start with Maximize Conversions to accumulate signal, move to target CPA (tCPA) or target ROAS (tROAS) only after stable conversion history exists, and set tROAS 20% below your historical average to start. Forcing a target before the system has learned destroys performance.[3]
  • Measurement requires more than last-click CPA. Demand Gen operates upstream of purchase intent, so engaged views, assisted conversions, view-through conversions, and Attributed Branded Searches must be included in any honest performance evaluation.[4]
  • Structure drives learning. One clear goal per campaign, one primary audience strategy per ad group, and minimal edits during the 14-day learning window are non-negotiable for giving Google’s AI enough clean signal to optimize.[3]

2. Benchmarks and Numbers at a Glance

Metric Typical range or threshold Applies when Source
CTR — all Demand Gen placements 0.50%–2.00% Vendor claim across mixed placement data; US accounts; sample size not stated [5]
CTR — YouTube placements only 1.20%–2.10% Vendor claim; YouTube in-feed and in-stream inventory; sample size not stated [5]
CTR — Discover placements only 1.50%–2.80% Vendor claim; Discover feed inventory; sample size not stated [5]
CPM — all Demand Gen placements $2.80–$5.20 Vendor claim; blended across YouTube, Discover, Gmail; US; sample size not stated [5]
CPM — YouTube placements only $4.50–$8.00 Vendor claim; YouTube inventory; consistently higher than Discover CPM; sample size not stated [5]
CPM — Discover placements only $3.20–$6.50 Vendor claim; Discover feed inventory; sample size not stated [5]
CPC — YouTube $0.55–$1.20 Vendor claim; YouTube clicks; sample size not stated [5]
CPC — Discover $0.45–$0.95 Vendor claim; Discover feed clicks; sample size not stated [5]
Ecommerce purchase CPA $24–$42 Vendor claim; ecommerce accounts; US; sample size not stated [4]
Ecommerce purchase ROAS 3.8×–5.2× Vendor claim; ecommerce accounts using tROAS bidding; US; sample size not stated [4]
Ecommerce CTR 0.50%–1.20% Vendor claim; ecommerce vertical specifically; slightly below blended benchmark [4]
Lookalike minimum seed size 100 active matched users (historical guidance) Vendor claim; Google Support as of July 2026 indicates this threshold may no longer apply as a hard constraint—use the most conservative figure until Google clarifies formally[1] [4]
Lookalike — narrow similarity tier ~2.5% of geo population (historical slider value) Vendor claim; Google Support July 2026 states slider values no longer function as hard targeting constraints; treat as directional only[1] [4]
Lookalike — balanced similarity tier ~5.0% of geo population (historical slider value) Same caveat as narrow tier above[1] [4]
Lookalike — broad similarity tier ~10.0% of geo population (historical slider value) Same caveat as narrow tier above[1] [4]
AI-expanded lookalike incremental conversions +8%–14% conversion volume at equivalent or slightly lower CPA Vendor claim; early test data from AI-expanded lookalikes; sample size not stated [6]
tROAS starting point vs. historical ROAS Set target 20% below historical average ROAS Google’s official value-based bidding guidance; applies when switching to tROAS for the first time [3]
Learning period — no edits window 14 days minimum Google’s official guidance; applies after any switch to value-based bidding; avoid structural changes during this window [3]
Budget rule of thumb for target strategies ~10× target CPA per day Vendor claim and Google-facing guidance; applies when using tCPA; ensures enough daily budget for the system to learn [1]

3. What Demand Gen Is and When to Use It

Demand Gen is Google’s AI-powered visual campaign type designed to surface new demand across YouTube (in-feed, in-stream, Shorts), Google Discover, and Gmail. It replaced Discovery campaigns and absorbed a portion of the legacy Display network’s function following Google’s migration of Display into Demand Gen.[7] Unlike Search, which harvests existing intent, Demand Gen is built to create intent by reaching users who match behavioral and audience profiles before they search for a product or category.

Demand Gen is the right choice when at least one of these conditions is true:

  • The account needs to grow the top of funnel because branded Search volume is stagnating and there is no new demand entering the category.
  • The creative assets are strong visually (video and high-quality imagery) and the team can maintain an ongoing creative rotation.
  • The product or offer benefits from contextual, feed-based discovery rather than query-triggered intent—fashion, home goods, travel, consumer apps, and subscription services are natural fits.
  • First-party audience lists are large enough to support lookalike expansion (historically 100 matched users, though Google’s July 2026 update suggests this threshold has changed—see Section 6).[1]
  • The measurement setup can accommodate view-through and assisted attribution rather than relying solely on last-click CPA.[4]

Demand Gen is not the right primary campaign when the account has no conversion tracking in place, when the daily budget is too small to exit the learning phase, or when the goal is purely branded query defense—Search handles that better.

Worked example

Deciding whether Demand Gen fits a new ecommerce account

  • Setup: A Chicago-based direct-to-consumer kitchenware brand spending $8,000/month on Google Ads, with 85% of budget currently in Search. Branded search impression share is 94% and Search clicks have plateaued for three consecutive months. The brand has 4,200 Customer Match records from email subscribers and a product catalog with lifestyle video assets in 16:9 ratio.
  • Numbers: $8,000/month total budget. Branded Search IS = 94%—essentially maxed out. Customer Match list = 4,200 records; well above the conservative 100-user seed threshold.[4] Benchmark ecommerce purchase CPA for Demand Gen = $24–$42.[4] To run tCPA at $35 target, daily budget needed = 10 × $35 = $350/day = $10,500/month. A dedicated Demand Gen budget of $3,500/month ($117/day) is feasible as a test but is below the 10× rule—use Maximize Conversions first.
  • Decision: Launch one Demand Gen campaign with a $117/day budget, Maximize Conversions bid strategy, and Customer Match + lookalike audience. Do not use tCPA until at least 30 conversions have accumulated over a stable 14-day window.
  • Why: Branded Search is saturated at 94% IS, so incremental volume must come from upper-funnel discovery; the 4,200-record Customer Match list clears the conservative seed threshold for lookalike expansion.[4]

4. Campaign Setup and Goals

A well-structured Demand Gen campaign starts with verified conversion tracking and a single, unambiguous campaign goal. Google’s guidance is explicit: conversion tracking must be confirmed working before campaign launch, not after.[8] Without it, bidding is flying blind and lookalike signals degrade over time.

Campaign structure rules

  • One goal per campaign. Mixing prospecting and remarketing in the same campaign splits the learning signal and makes performance attribution unreliable. Use separate campaigns for each objective.[1]
  • One primary audience strategy per ad group. Do not stack Customer Match, lookalike, and custom segments in a single ad group. Separate them so Google can learn which signal drives conversions and so you can make informed budget decisions.[1]
  • Verify the Google tag before launch. Use Google Tag Assistant to confirm the conversion action fires correctly on the actual conversion page. Sitewide tagging and, where applicable, the Google Tag Gateway improve signal completeness for bidding.[8]
  • Use Performance Planner before budget changes. Google’s February 2026 guidance explicitly recommends Performance Planner to model the impact of budget and bid changes before applying them, reducing the risk of disrupting the learning phase.[1]

Goal and conversion action selection

Select conversion actions that reflect a meaningful business outcome. For ecommerce, that is a purchase. For lead gen, that is a qualified form submission or phone call. If conversion volume is too low to train the bidding model (fewer than 30 conversions in 30 days is a common practitioner threshold), temporarily add a higher-funnel micro-conversion—add to cart, page view with duration—as a secondary optimization signal while the campaign accumulates data.[8]

Worked example

Setting up conversion tracking for a B2C lead gen account

  • Setup: A Tampa, FL home solar installer running Demand Gen for the first time. Primary conversion = quote form submission, which historically fires 18 times per month across all campaigns combined. The team wants to use tCPA immediately.
  • Numbers: 18 form submissions per month = 0.6/day on average. tCPA requires roughly 30+ conversions over a stable 14-day window to train effectively.[3] 0.6/day × 14 days = 8.4 conversions—far below threshold. Benchmark blended CTR = 0.50%–2.00%.[5] If the campaign runs at 0.75% CTR and the site converts at 3%, the account needs ~1,111 clicks to get 33 conversions; at a Discover CPC of $0.45–$0.95, that is $500–$1,055 in spend just to exit learning.
  • Decision: Add “Schedule a Call” button click (a higher-frequency micro-conversion firing ~90 times/month) as the primary optimization event for the first 30 days. Switch to form submission as the sole conversion action once Demand Gen alone is generating 30+ form submissions per 30-day period.
  • Why: Forcing tCPA on 18 total monthly conversions across all campaigns means the Demand Gen campaign alone will never accumulate enough signal to exit learning; the micro-conversion provides 5× more signal while remaining directionally correlated with the actual business outcome.[8]

Campaign setup checklist

  • Google Tag verified and firing on the correct conversion page before go-live.[8]
  • One campaign per goal (prospecting, remarketing, or new customer acquisition).[1]
  • One primary audience per ad group.[1]
  • At least one ad with Excellent Ad Strength in each campaign.[2]
  • Minimum 3 unique assets per aspect ratio (vertical 9:16, square 1:1, landscape 16:9) per ad group.[2]
  • Bid strategy matched to data maturity: Maximize Conversions first, tCPA/tROAS after 30+ stable conversions.[3]
  • Performance Planner reviewed before any budget change exceeding 20%.[1]

Done for you

Everything on this page, done for you before you launch

We build the campaign properly the first time — so you are not unpicking it three months from now.

✓  Keywords validated against real search volume, grouped by intent

✓  Responsive search ads written to the asset counts that actually perform

✓  A negative keyword list built before you spend, not after

✓  Delivered in 24 hours as a spreadsheet you keep

See what you get — $129

Human-reviewed · No subscription · We never touch your Google Ads account

5. Creative and Ad Formats

Creative is the single highest-leverage variable in Demand Gen. Because Google’s AI automatically allocates spend across YouTube, Discover, and Gmail based on predicted performance, the breadth and quality of the creative pool directly determines how efficiently that allocation works. A campaign with one hero image and one video in a single aspect ratio will underserve placements and leave inventory untapped.[2]

Asset requirements by format

Asset type Minimum recommended Aspect ratios required Key quality rules
Images 3 unique assets per ratio 9:16 vertical, 1:1 square, 16:9 landscape Minimal text overlay, strong lighting, lifestyle or product imagery[8]
Videos At least 1; ideally 3+ across ratios 9:16 for Shorts, 16:9 for in-stream/in-feed Hook in first 5 seconds; brand mention early; authentic over polished[2]
Headlines 3–5 unique headlines per ad N/A Varies by placement; write to communicate offer without image context[9]
Descriptions 2–3 unique descriptions per ad N/A Concise; test benefit-led vs. feature-led variants[9]

Ad Strength and the “rule of three”

Google’s creative guidance introduces the “rule of three”: include at least 3 unique assets for each aspect ratio in every ad group. The goal is to have at least one ad per campaign rated “Excellent” for Ad Strength, which Google correlates with stronger performance and broader reach across visual inventory.[2] Ad Strength ratings are visible in the Ads section of the campaign and update as assets are added or replaced.

Placement-specific creative priorities

  • YouTube (in-feed, in-stream, Shorts): Video is essential. Shorts specifically requires 9:16 vertical format. Google added 9:16 vertical image ads for Shorts in 2025–2026, so include vertical images even if video is unavailable.[7] Creator-style, authentic video consistently outperforms high-production spots on YouTube inventory.[6]
  • Discover: Highly visual feed. Use lifestyle imagery with minimal text. Google’s own spec guidance calls for clean product shots and effective lighting. Discover CTR benchmarks (1.50%–2.80%) are the highest of any Demand Gen placement, so creative quality here has an outsized ROI impact.[5]
  • Gmail: Static creative with tightly written copy performs best. The Gmail format is more text-adjacent than YouTube or Discover, so headlines and descriptions carry more weight. Keep imagery clean and the offer explicit in the first headline.[4]

Multimodal Video Creation in Asset Studio

As of August 2026, Multimodal Video Creation in Asset Studio is generally available in the US. Advertisers can move from a text or image storyboard to both horizontal (16:9) and vertical (9:16) video assets within a single Asset Studio workflow, reducing the production barrier for teams without dedicated video resources.[7]

Worked example

Building a compliant creative set for a DTC apparel brand

  • Setup: A Portland, OR DTC activewear brand launching its first Demand Gen campaign for a September 2026 fall collection drop. Budget is $5,000/month. The team has 6 lifestyle product photos (all 16:9 landscape) and one 30-second brand video (16:9 only). Ad Strength reads “Poor.”
  • Numbers: Current assets: 6 images (1 ratio only) + 1 video (1 ratio only). Required per Google’s rule of three: 3 unique assets × 3 ratios = 9 image assets minimum.[2] Missing: 3 vertical (9:16) images, 3 square (1:1) images, and a 9:16 vertical video for Shorts. Discover CTR benchmark = 1.50%–2.80%.[5] At $5,000/month and a blended CPM of $3.20–$5.20, expected impressions = 961,538–1,562,500. With 1.50% CTR (low end), that is 14,423–23,437 clicks/month. Adding missing ratios to reach Excellent Ad Strength is the single highest-leverage action before increasing budget.
  • Decision: Before launch, use Asset Studio’s Multimodal Video Creation to generate a 9:16 vertical version of the existing 16:9 video, and crop the 6 landscape photos into 9:16 and 1:1 variants. This reaches 9 images across 3 ratios + 2 videos across 2 ratios. Recheck Ad Strength; target “Excellent” before setting the campaign live.[2]
  • Why: Google requires at least 3 unique assets per aspect ratio to unlock the full creative rotation across placements; without vertical and square assets, Shorts and portions of Discover inventory are effectively unreachable.[2]

Worked example

Prioritizing creative refresh cadence on a fixed monthly budget

  • Setup: A San Francisco B2C software subscription brand spending $12,000/month on Demand Gen. The campaign has been live for 90 days. CTR has dropped from 1.8% in month 1 to 0.9% in month 3—consistent with creative fatigue. The creative set has 3 images per ratio and 1 video (all 16:9).
  • Numbers: CTR decline: 1.8% → 0.9% = 50% drop over 60 days. Benchmark YouTube CTR = 1.20%–2.10%; 0.9% is now below the lower bound.[5] At blended CPM of $4.00 and $12,000/month spend, impressions = 3,000,000/month. At 0.9% CTR = 27,000 clicks vs. 54,000 clicks at 1.8%—a loss of 27,000 clicks/month at $0.55–$1.20 CPC (YouTube) effectively wasted ~$14,850–$32,400 in opportunity cost.
  • Decision: Introduce 3 new image creatives (1 per ratio) and a second 16:9 video with a different hook within the next 7 days. Pause the 2 lowest-performing existing images by impression-weighted CTR. Set a calendar reminder to audit creative performance every 30 days going forward.
  • Why: CTR falling below the lower benchmark bound (1.20% for YouTube)[5] is the signal to refresh; adding net-new assets without removing low performers inflates the pool without retiring fatigue drivers.

6. Audiences and Lookalike Segments

Audience strategy in Demand Gen follows a strict hierarchy: first-party data first, lookalike expansion second, custom segments as a supplement. Google’s February 2026 guidance consolidates this into three core audience levers—optimized targeting, lookalike segments, and new customer acquisition goals—as the recommended starting point for any Demand Gen setup.[1]

First-party audiences

Customer Match, website visitor lists, app users, and offline conversion imports are the highest-quality signals available in Demand Gen. Use them in two ways: as direct remarketing audiences and as seed lists for lookalike expansion. Seed lookalikes from your best customers or highest-value converters, not from broad site visitor lists.[1] When the campaign goal is new customer acquisition, exclude existing customers explicitly using a Customer Match exclusion list to prevent spend on users who would convert anyway.[1]

Lookalike segments

Lookalike segments remain one of the most powerful scaling tools in Demand Gen. Google’s AI builds lookalike audiences from your seed lists using behavioral and interest signals across Google’s logged-in user base.[4]

An important discrepancy exists in the research: earlier vendor guidance cited a hard minimum of 100 active matched users for seed list eligibility and fixed reach tiers of 2.5%, 5.0%, and 10.0% of the geographic population.[4] However, Google’s official support page updated as of July 2026 indicates the 100-user threshold and the fixed tier percentages no longer function as hard targeting constraints.[1] Recommended conservative approach: continue building seed lists to at least 100 matched users and treat the similarity tier slider as directional guidance rather than a precise audience size control until Google provides formal replacement documentation.

Early test data suggests AI-expanded lookalikes drive 8%–14% more conversion volume at equivalent or slightly lower CPA compared to fixed-tier lookalikes.[6] Treat this as directional, not guaranteed, given the vendor-claim provenance and unknown sample size.

Custom segments

Custom segments—built from search terms users have entered on Google, apps they have used, or websites they have visited—are useful for capturing intent-based prospecting signals that first-party data cannot express. Use them as a secondary layer in prospecting ad groups, not as a replacement for Customer Match or lookalike seeds. Keep each custom segment focused on a single theme; combining unrelated keywords or interests in one segment dilutes the signal.[1]

Optimized targeting and new customer acquisition

Optimized targeting allows Google to expand delivery beyond your defined audience when conversion probability is higher outside your segment. Enable it for prospecting campaigns where the goal is broad new customer acquisition. Disable it for tight remarketing ad groups where you specifically want to reach only your defined list.[1]

New Customer Acquisition goal mode instructs Demand Gen to prioritize impressions and conversions among users who do not appear in your existing customer lists. It is most effective when paired with a clean, up-to-date Customer Match exclusion list.[1]

Worked example

Structuring a three-ad-group lookalike prospecting campaign

  • Setup: A Minneapolis B2C fitness equipment brand with three Customer Match lists: (A) 8,200 purchasers in the last 12 months, (B) 14,500 email subscribers who have never purchased, and (C) 3,100 repeat purchasers (2+ orders). The team wants to maximize new customer acquisition on a $9,000/month Demand Gen budget.
  • Numbers: List A = 8,200 purchasers; List C = 3,100 repeat purchasers. Conservative seed threshold = 100 matched users—all three lists clear it comfortably.[4] If AI-expanded lookalikes drive +11% conversion volume (midpoint of 8%–14% range)[6] vs. fixed-tier lookalikes, and the baseline is 90 conversions/month, that equals ~10 incremental conversions/month at equivalent CPA. Budget allocation: Ad Group 1 (lookalike from List C repeat purchasers) = $4,000/month; Ad Group 2 (lookalike from List A purchasers) = $3,500/month; Ad Group 3 (custom segment, category intent keywords) = $1,500/month.
  • Decision: Create 3 separate ad groups, one per seed strategy. Enable New Customer Acquisition goal at campaign level. Add List A + List B + List C as exclusions on all three ad groups so no known customers receive prospecting impressions. Set Maximize Conversions for all three ad groups for the first 30 days before evaluating tCPA eligibility.
  • Why: Separating seed types into distinct ad groups lets Google optimize each signal independently and lets the team attribute performance to the correct audience lever; mixing all three seeds into one ad group would obscure which seed is actually driving results.[1]

Not ready to change anything today?

We post the sharpest bits of these guides — what a change is worth, how long it takes to show up, and when to leave it alone. Short enough to read on your phone.

Follow on LinkedIn Follow on Instagram

7. Bidding Strategies

Bid strategy selection in Demand Gen is a function of two variables: the campaign’s primary objective and the volume and stability of conversion data available. Using the wrong strategy for the data maturity stage is one of the most common and costly errors in Demand Gen account management.[3]

Bid strategy selection by scenario

Scenario Recommended strategy When to move to next stage
New campaign, fewer than 30 conversions in 30 days Maximize Conversions After 30+ conversions over a stable 14-day window
Established campaign, clear CPA target, 30+ conversions/month tCPA When value data is available and ROAS target is defined
Ecommerce, revenue optimization, clear ROAS target tROAS set 20% below historical ROAS Raise tROAS by no more than 10%–15% increments after 14-day stable periods
Awareness / top-of-funnel only Maximize clicks or CPM-based where available When downstream conversion data accumulates sufficiently

Budget sizing for target strategies

Google’s guidance and practitioner consensus align on a daily budget of approximately 10× the tCPA target to give the bidding system enough room to explore and optimize.[1] A campaign with a $40 tCPA should have at least a $400/day ($12,000/month) budget. Campaigns below this threshold will frequently stall in learning or oscillate between learning and active status.

The learning period

After switching to any value-based bid strategy (tCPA or tROAS), Google explicitly instructs advertisers to leave the campaign unchanged for the first 14 days.[3] This means no audience changes, no budget changes exceeding ~20%, no creative swaps, and no bid target adjustments. Use Performance Planner to simulate changes before applying them.[1] Evaluating performance day-by-day during the learning window and making reactive changes is one of the fastest ways to permanently impair a campaign’s ability to optimize.

tROAS setup guidance

Google’s value-based bidding guidance specifies starting tROAS at 20% below the campaign’s historical average ROAS.[3] If the historical ROAS is 4.5×, the starting tROAS should be set to 3.6× (4.5 × 0.80 = 3.6). After a stable 14-day learning period, incrementally raise the target by 10%–15% at a time. Raising tROAS aggressively in a single step typically triggers a new learning period and often reduces conversion volume sharply.

Worked example

Transitioning from Maximize Conversions to tROAS on an ecommerce campaign

  • Setup: A Denver outdoor gear ecommerce brand running a Demand Gen campaign for 45 days on Maximize Conversions. The campaign has generated 62 purchases over the last 30 days at an average order value of $148 and a ROAS of 4.8×. The team wants to switch to tROAS to improve efficiency heading into Q4 2026.
  • Numbers: Historical ROAS = 4.8×. Starting tROAS = 4.8 × 0.80 = 3.84× (round to 3.8×). 62 conversions in 30 days = 2.1/day—well above the 30-conversion threshold for target bidding.[3] Benchmark ecommerce ROAS range = 3.8×–5.2×.[4] Starting tROAS of 3.8× sits at the lower bound of the benchmark range, giving the system room to deliver volume. Daily budget = $400 (current); target CPA implied = $148 ÷ 4.8 = $30.83; 10× rule = $308/day minimum—current $400/day clears the threshold comfortably.[1]
  • Decision: Switch bid strategy from Maximize Conversions to tROAS, set target to 3.84× (entered as 384% in Google Ads UI), leave all other settings unchanged for exactly 14 days, then evaluate before any further adjustment.[3]
  • Why: Google’s value-based bidding guidance specifies starting tROAS 20% below historical average ROAS to ensure the system has sufficient room to win auctions while transitioning; entering at the historical ROAS of 4.8× immediately risks volume collapse during learning.[3]

8. Placements: YouTube, Discover, and Gmail

Demand Gen automatically allocates spend across YouTube, Discover, and Gmail based on predicted performance. Advertisers cannot manually set percentage splits between placements, but they can use channel controls to include or exclude specific surfaces.[7] The practical implication is that creative must be prepared for all three environments because Google will use whichever placement it predicts will deliver the best result for the bid and audience at any given moment.[10]

YouTube

YouTube receives the highest CPMs within Demand Gen ($4.50–$8.00)[5] but also delivers the highest intent signals. Demand Gen runs across YouTube in-feed (next to watch page recommendations), in-stream (before and during videos), and Shorts (the vertical, short-form feed). CTR benchmarks for YouTube (1.20%–2.10%)[5] sit above the blended average, reflecting the higher engagement of an active viewing environment.

As of August 2026, Google is testing YouTube messaging placements that allow viewers to initiate conversations with brands on messaging apps directly from a Demand Gen ad on YouTube.[7] This feature is in test; do not build a core conversion strategy around it until it reaches general availability.

For Shorts specifically, 9:16 vertical format is required for video assets. Google also added 9:16 vertical image ads for Shorts, making it possible to participate in Shorts inventory without a dedicated vertical video if a vertical image is uploaded.[7]

Discover

Discover delivers the highest CTR of any Demand Gen placement (1.50%–2.80%)[5] at a lower CPM ($3.20–$6.50)[5] than YouTube, making it cost-efficient for driving clicks. The Discover feed is consumed passively by users browsing content recommendations, so creative must stop the scroll visually rather than relying on audio or motion. Use full-bleed lifestyle imagery, minimal or no text overlay, and clear product framing.[8]

Gmail

Gmail ads appear as collapsed promotions in the Gmail Promotions tab and expand into a full-page ad experience on click. Gmail CPM and CPC data was not surfaced in the research for this article—use the blended benchmark ($2.80–$5.20 CPM)[5] as a conservative proxy. Gmail responds better to clear, benefit-led copy and a single focused call to action. Because the placement is email-adjacent, users’ expectations are informational; ads that mirror an email’s structure (subject line as headline, body as description) tend to outperform purely visual treatments.[4]

Channel controls

Channel controls, available as of 2026, allow advertisers to enable or disable delivery on individual Demand Gen surfaces: YouTube, Discover, Gmail, and Google Display Network.[7] Use channel controls to exclude a placement only when you have strong creative-fit or brand-safety reasons to do so, not as a routine optimization lever. Restricting placements reduces the AI’s inventory access and typically increases CPM and CPA by limiting supply.

Worked example

Using channel controls after a brand-safety incident on Gmail

  • Setup: A New York financial services brand running Demand Gen for customer acquisition. Compliance reviews the Gmail ad experience in October 2026 and determines that the expanded Gmail ad format displays fine-print disclosures in a way that does not meet FINRA guidelines. The brand needs to pause Gmail delivery while the compliance team approves a revised creative treatment.
  • Numbers: Current budget = $15,000/month. Gmail’s estimated share of Demand Gen delivery at blended CPM $3.50 (conservative mid-point)[5]: if Gmail accounts for ~15% of impressions, that is ~642,857 Gmail impressions/month. Excluding Gmail reduces available supply by ~15%, which may increase blended CPM by approximately 5%–10% based on reduced inventory access—a cost increase of $750–$1,500/month. This is the quantified trade-off for compliance.
  • Decision: Navigate to campaign settings → Channel controls, deselect Gmail, save. Set a task for November 15, 2026 to re-enable Gmail once the compliance-approved creative is uploaded and verified.
  • Why: Channel controls are the correct tool for placement-specific brand safety or compliance actions; adjusting bids or budgets does not exclude a placement, it only reduces its relative share.[7]

“So what do I actually have to hand over?”

Nothing. No access to your Google Ads account, no login, no agency link request. You tell us the business and the offer; we send back a spreadsheet.

You import it yourself, or hand it to whoever runs the account. It is yours to keep and re-use — there is no ongoing fee and nothing to cancel.

How it works — $129

9. Measurement and View-Through Attribution

Demand Gen’s measurement challenge is structural: the format is designed to reach users before they have expressed search intent, so last-click attribution systematically undercounts its contribution. A user who sees a Demand Gen ad on YouTube on Monday, searches the brand on Thursday, and converts via a Search ad on Friday will appear as a Search-only conversion in last-click reports. The Demand Gen impression that triggered the consideration is invisible.[4]

Building a complete measurement stack

A robust Demand Gen measurement setup includes all of the following, evaluated together rather than individually:

  • Last-click conversions: Standard conversion reporting; necessary but insufficient for Demand Gen evaluation.
  • Assisted conversions: Shows Demand Gen’s role in conversion paths where another campaign received the last click. Found in the Attribution reports section of Google Ads.
  • View-through conversions: Counts conversions where the user saw a Demand Gen ad but did not click, then converted within the attribution window. Use a conservative view-through window (1–7 days) to avoid over-attribution; the default 30-day window is too generous for most accounts.[4]
  • Engaged views: A YouTube-specific metric counting views where the user watched at least 10 seconds or completed a shorter video. Engaged view rate is a leading indicator of creative quality before conversion data matures.[4]
  • Attributed Branded Searches: A measurement feature (generally available as of January 2026) that surfaces the volume of branded searches on Google and YouTube generated by your Demand Gen campaigns. This is the most direct evidence of demand creation rather than demand harvesting.[7]

Data quality requirements

Bidding quality and measurement accuracy both depend on conversion data completeness. Google’s 2026 guidance consistently emphasizes sitewide tagging (Google tag on every page), consideration of the Google Tag Gateway for improved signal durability in privacy-restricted environments, and offline conversion imports via Data Manager for lead gen accounts that close business offline.[8] Each of these reduces the gap between observed and actual conversions, which directly improves tCPA and tROAS model accuracy.

Attribution window stability

Do not change conversion windows mid-test. Changing the attribution window while a campaign is in a learning phase or while a performance evaluation is in progress distorts both bidding signals and comparative reporting. Fix the window before launch and document the date any change is made.[3]

Worked example

Evaluating Demand Gen using Attributed Branded Searches alongside last-click CPA

  • Setup: An Austin, TX consumer app brand has run Demand Gen for 60 days at $6,000/month. Last-click CPA from the Demand Gen campaign is $58, above the $45 tCPA target. The team is considering pausing the campaign. Before pausing, they pull Attributed Branded Searches and the Attribution paths report.
  • Numbers: Last-click Demand Gen CPA = $58 vs. $45 target—29% above goal. Attributed Branded Searches report shows 1,240 incremental branded searches over the 60-day period attributed to Demand Gen exposure. The brand’s Search campaign converts branded queries at a 12% conversion rate with an average CPA of $18. Implied incremental Search conversions from Demand Gen: 1,240 × 12% = 148.8, call it 149 conversions. Cost of those 149 conversions attributed back to Demand Gen spend: 149 × $18 = $2,682. Demand Gen spend over 60 days = $12,000. Total attributed conversions: direct last-click conversions (say 207 at $58 CPA = $12,006 spend) + 149 influenced Search conversions = 356 total. Blended CPA: $12,000 ÷ 356 = $33.71—below the $45 target.
  • Decision: Do not pause. Instead, present the blended CPA of $33.71 (which includes attributed branded search lift) to stakeholders and continue the campaign. Set up a monthly Attributed Branded Searches pull as a standing KPI.
  • Why: Evaluating Demand Gen on last-click CPA alone systematically excludes its demand-creation effect on branded Search; the Attributed Branded Searches feature is the correct tool to quantify that upstream contribution.[7]

10. Demand Gen vs. Performance Max vs. Video Campaigns

These three campaign types overlap significantly in inventory access but differ in objective, control, and optimization mechanism. Choosing the wrong type for the goal wastes budget and obscures performance data. No direct cost-and-conversion comparison data between the three types was available in peer-reviewed or independently audited research as of September 2026; the distinctions below are based on Google’s documentation and practitioner guidance.[4]

Feature comparison

Feature Demand Gen Performance Max Standard Video campaign
Primary purpose Upper-to-mid funnel demand creation; visual discovery Full-funnel conversion across all Google inventory Brand awareness or video-specific reach/engagement
Placements YouTube, Discover, Gmail, GDN All Google inventory including Search, Shopping, Display, YouTube, Maps YouTube only (in-stream, in-feed, Shorts, bumper)
Audience control High: Customer Match, lookalikes, custom segments, channel controls Low: signal suggestions only; AI controls delivery High: standard audience targeting and exclusions
Creative requirements Images + video across multiple ratios All asset types fed as inputs; AI assembles formats Video only
Bid strategies available Maximize Conversions, tCPA, tROAS, Maximize Clicks Maximize Conversions, Maximize Conversion Value, tCPA, tROAS CPV, CPM, tCPA, Maximize Conversions
Learning period (no-edit window) 14 days for value-based strategies[3] Typically 14–28 days; longer due to broader inventory Standard; shorter due to narrower inventory
Transparency Placement and audience breakdowns available Limited; asset group reporting only Full placement and demographic reporting
Best fit Accounts needing mid-funnel visual discovery with audience control Accounts wanting Google AI to maximize conversions across all surfaces with minimal manual control Accounts with video-only assets and a YouTube-specific reach or engagement goal

When to use each type

Use Demand Gen when the goal is to create demand among users who match a defined audience profile across visual surfaces, you want to retain control over audience targeting, and you have both image and video assets. Demand Gen is the correct choice when the account is targeting a specific lookalike or Customer Match population and needs placement-level reporting.[4]

Use Performance Max when the priority is maximizing conversion volume across all Google surfaces with minimal manual management, and the account is mature enough (typically 30–50+ conversions/month) for the AI to optimize effectively. Note that Performance Max and Demand Gen can coexist in the same account but may compete in the same auctions; segment their goals and audiences clearly.[4]

Use Standard Video campaigns when the goal is YouTube-specific brand awareness, reach frequency control, or CPV-based engagement, and the account does not have image assets available. Standard Video campaigns offer the most control over YouTube-specific formats and targeting.[4]

Worked example

Choosing between Demand Gen and Performance Max for a mid-market B2C brand

  • Setup: A Seattle consumer electronics brand generating 140 purchases/month across all Google campaigns at a blended CPA of $37. The marketing director wants to scale to 200 purchases/month by Q4 2026 without increasing CPA above $42. The account already runs Shopping and Search. The team has 9 creative assets (3 per ratio) and one 15-second product video.
  • Numbers: Current purchases = 140/month; target = 200/month = +43% volume. Current CPA = $37; CPA ceiling = $42 (+13.5% tolerance). 140 conversions/month ÷ 30 days = 4.7/day—well above the 30-conversion threshold for tCPA on Demand Gen.[3] At tCPA of $40 and 10× budget rule, required daily budget = $400/day = $12,000/month for Demand Gen. The team has $10,000/month available for new spend. Performance Max would also access Search and Shopping inventory currently covered by existing campaigns, creating auction competition and attribution confusion.
  • Decision: Launch Demand Gen (not Performance Max) at $10,000/month ($333/day) with tCPA = $40, targeting lookalike from the 1,400-record purchaser Customer Match list. Do not launch Performance Max, as it would overlap with existing Search and Shopping campaigns on the same conversion actions, making attribution and incrementality impossible to interpret.
  • Why: Demand Gen provides audience-controlled expansion into YouTube/Discover/Gmail without competing with existing Search and Shopping campaigns in the same auctions; Performance Max would cannibalize existing campaign traffic and obscure incremental contribution.[4]

11. Common Mistakes to Avoid

The following mistakes appear consistently in practitioner accounts and Google’s own guidance. Each one is specific, actionable to fix, and directly tied to performance degradation.

Structural mistakes

  • Launching without verified conversion tracking. The most fundamental error. If the Google tag is not confirmed firing on the actual conversion page before launch, every subsequent optimization signal is corrupted. Use Google Tag Assistant to verify before setting the campaign live.[8]
  • Mixing prospecting and remarketing in the same campaign. Blending audiences with different funnel positions in one campaign prevents Google from optimizing each signal correctly and makes performance analysis meaningless. Use separate campaigns.[1]
  • Forcing tCPA or tROAS before sufficient conversion history. Applying a target strategy with fewer than 30 stable conversions per 30-day period causes the bidding system to oscillate, frequently triggering repeated learning phases and delivering no consistent volume.[3]
  • Making changes during the 14-day learning window. Any significant structural change—new audiences, bid target change, budget cut exceeding ~20%—resets the learning clock. Reactive day-by-day management during learning is among the most costly account management errors.[3]

Creative mistakes

  • Uploading assets in only one aspect ratio. A campaign with only 16:9 landscape images cannot serve Shorts inventory (requires 9:16) or square-format placements. This leaves significant inventory unreachable and raises effective CPM by restricting supply.[2]
  • Heavy text overlay on images. Google’s own creative guidance explicitly flags text-heavy images as a quality issue for Discover and YouTube. Images with large text blocks score lower in Ad Strength and underperform in feed placements.[8]
  • Running the same creative for more than 60 days without a refresh audit. Creative fatigue typically manifests as a CTR decline of 30%–50% from the first-month baseline. Audit creative performance monthly and retire the bottom-performing assets by impression-weighted CTR.[2]

Measurement mistakes

  • Evaluating Demand Gen solely on last-click CPA. Last-click attribution structurally undercounts Demand Gen’s contribution by missing view-through influence, assisted conversions, and downstream branded search lift. Always include Attributed Branded Searches and assisted conversion data in any performance review.[4]
  • Using a 30-day view-through conversion window without justification. The default 30-day view-through window is generous and inflates conversion counts on Demand Gen campaigns. Use a 1–7-day view-through window as a conservative starting point and document the window so stakeholders understand the attribution model.[4]
  • Seeding lookalikes from low-quality or broad lists. Building lookalikes from generic site visitor lists rather than from high-value purchasers or converters produces audiences that are statistically similar to your worst customers, not your best. Always seed from the highest-quality sub-segment of your first-party data.[1]

Worked example

Diagnosing a campaign stuck in repeated learning phases

  • Setup: A Boston SaaS brand running Demand Gen for B2B lead gen at $4,000/month ($133/day) with tCPA set to $75. The campaign has been live for 6 weeks and has generated only 11 form submissions. The account manager has changed the tCPA target 4 times (from $75 to $60 to $90 to $75 to $65) over those 6 weeks in response to day-by-day CPA fluctuations. The campaign shows “Learning” status in the bid strategy report almost continuously.
  • Numbers: 11 conversions in 6 weeks = 1.83/week = 0.26/day. Required for stable tCPA: 30+ conversions over 14 days = 2.1/day minimum—current rate is 92% below threshold. Daily budget = $133; 10× tCPA rule at $75 target = $750/day required—current budget is 82% below the recommended minimum.[1] 4 bid target changes in 42 days = a new learning clock reset approximately every 10 days, ensuring the campaign never exits learning.
  • Decision: (1) Switch immediately from tCPA to Maximize Conversions. (2) Add “Request a Demo” button click (~45 fires/month estimated) as a secondary optimization event to boost signal. (3) Increase budget to $300/day if possible, or accept that tCPA is not achievable at $133/day until conversion volume triples. (4) Make no further bid or audience changes for 21 days and evaluate again.
  • Why: A tCPA campaign generating 0.26 conversions/day will never exit learning; 4 bid changes in 42 days reset the 14-day no-edit window each time, compounding the problem—switching to Maximize Conversions removes the learning dependency on a conversion volume the account cannot currently support.[3]

12. What Changed Recently (Last 30 Days)

The updates below reflect Google’s confirmed Demand Gen product changes as of August–September 2026. All changes are sourced from Google’s official blog, help center, and product announcement channels. Where a feature is in limited testing rather than general availability, that status is noted.

YouTube Messaging Placements (Testing)

Google is actively testing the ability for viewers to initiate a conversation with a brand on a messaging app—such as WhatsApp or Google Messages—directly from a Demand Gen ad running on YouTube.[7] When a viewer taps the call-to-action overlay, the ad hands off to the advertiser’s chosen messaging channel without requiring the viewer to visit a landing page first. This is particularly relevant for service businesses, B2B lead gen, and any account where a direct conversation drives more qualified pipeline than a form fill.

At the time of writing, YouTube messaging placements are in a limited test, not generally available to all US advertisers. Do not build a Q4 2026 campaign plan around this feature without first confirming access in your Google Ads account or with your Google rep.[7]

Multimodal Video Creation in Asset Studio: Now Generally Available

Multimodal Video Creation inside Asset Studio moved from beta to general availability in August 2026.[7] The tool allows advertisers to go from a text prompt or a storyboard brief to both a horizontal (16:9) and a vertical (9:16) video asset within a single workflow. This matters for Demand Gen specifically because the campaign type distributes across YouTube in-feed, in-stream, and Shorts placements—all of which benefit from having purpose-built aspect ratios rather than algorithmically cropped versions of a single master cut.

Google’s creative guidance requires at least 3 unique assets per aspect ratio per ad group.[2] Multimodal Video Creation reduces the production cost of meeting that requirement for vertical formats, which previously required a separate edit from a video production team.

Worked example

Using Multimodal Video Creation to Meet the 3-Asset-Per-Ratio Requirement for a Shorts-Heavy Campaign

  • Setup: A Chicago fitness apparel account spending $8,000/month on a Demand Gen campaign targeting 25–44-year-old women in the continental US. The account currently has 2 landscape (16:9) video assets and 0 vertical (9:16) assets. Shorts inventory on YouTube is under-monetized as a result, because Google’s delivery system deprioritizes ad groups that cannot match the native aspect ratio of a placement.[2]
  • Numbers: YouTube Shorts CPM benchmark = $4.50–$8.00 (YouTube placement).[5] At a $8,000/month budget ($267/day) and a midpoint CPM of $6.25, the campaign is eligible for approximately 42,700 impressions/day across YouTube inventory. Without vertical assets, Shorts inventory—which Google estimates accounts for a material share of YouTube’s daily active surface area—is effectively unavailable. Adding 3 vertical (9:16) videos via Asset Studio costs $0 in incremental production spend using the new GA tool vs. an estimated $1,200–$2,400 for an external agency edit of 3 clips.
  • Decision: Use Asset Studio’s Multimodal Video Creation to generate 3 vertical (9:16) video assets from the existing 16:9 master cuts. Upload all 3 to the active ad group alongside the existing landscape assets. Set Asset Studio output to 9:16 at 1080×1920px, minimum 5 seconds, maximum 60 seconds per Google’s Shorts ad spec.[7]
  • Why: Google’s creative best-practice guidance requires at least 3 unique assets per aspect ratio; the account had 0 vertical assets, making it ineligible for Shorts delivery and leaving a meaningful share of YouTube CPM inventory untapped at the campaign’s existing $267/day spend level.[2]

Travel Feeds and Personalized Hotel Ads

Google expanded Demand Gen’s travel-specific capabilities in the August 2026 drop, adding the ability to surface local activities, real-time events, and personalized hotel properties to users whose signals indicate travel intent.[7] Advertisers in the travel vertical can now connect a Hotel Center feed to Demand Gen to generate dynamic video ads that display live hotel pricing, star ratings, and availability.[7] This was previously available in limited form as part of the January 2026 Demand Gen Drop’s Travel Feeds announcement.[11]

For non-travel advertisers, the practical implication is that product feed connectivity is expanding across more verticals and ad surfaces. Ecommerce and retail accounts should verify whether their Merchant Center feed is correctly linked to their Demand Gen campaigns, because Google is broadening the inventory surfaces on which feed-driven creative can appear.[7]

Shoppable Connected TV (CTV)

Demand Gen can now power shoppable ad experiences on connected TV screens via YouTube, enabling viewers watching YouTube on a television to browse and purchase products without switching to a second device.[11] The experience surfaces a product carousel beneath or alongside the video creative when a viewer shows engagement signals (pause, remote click, or extended dwell).

Shoppable CTV requires a linked Merchant Center feed and is currently most effective for accounts that already run Demand Gen with product feeds enabled on mobile and desktop placements. CPMs on CTV inventory tend to run at or above the YouTube placement benchmark of $4.50–$8.00,[5] so accounts should factor this into budget modeling when CTV delivery share increases.

Worked example

Budget Impact of CTV Impression Share Growth for a Direct-to-Consumer Kitchenware Account

  • Setup: A Minneapolis direct-to-consumer kitchenware account running a Demand Gen campaign at $5,000/month ($167/day) with a Merchant Center feed linked and shoppable CTV enabled after the August 2026 update. The account’s Google Ads placement report shows CTV now accounts for 18% of YouTube impressions served in the first 2 weeks after enabling the format.
  • Numbers: Total daily impressions at $167/day and a blended YouTube CPM of $6.25 (midpoint of $4.50–$8.00 benchmark[5]) = approximately 26,700 impressions/day. CTV share at 18% = 4,800 CTV impressions/day. If CTV CPM comes in at the top of the YouTube benchmark ($8.00), those 4,800 impressions cost $38.40/day vs. $30.00/day at the $6.25 blended rate—an incremental cost of $8.40/day, or approximately $252/month. At 20% CTV share and an $8.00 CPM, the effective blended CPM rises from $6.25 to $6.55, consuming the $5,000 budget roughly 4.6 days earlier per month.
  • Decision: Increase the monthly Demand Gen budget by $260 (from $5,000 to $5,260) to preserve impression volume after CTV delivery begins. Monitor the placement report weekly; if CTV share exceeds 25%, re-model at the $8.00 CPM ceiling to avoid unintended budget compression on Discover and Gmail placements.[5]
  • Why: CTV inventory within YouTube Demand Gen prices at the top of the $4.50–$8.00 YouTube CPM range; failure to account for the CPM premium when CTV share grows will compress total impression volume without a corresponding budget increase.[5]

Attributed Branded Searches

Google introduced the Attributed Branded Searches measurement feature, which shows the volume of branded search queries on Google and YouTube that are attributed to exposure from a Demand Gen campaign.[11] This metric is designed to help advertisers quantify the upper-funnel demand-creation effect of Demand Gen activity—specifically, whether ad exposure is causing people to subsequently search for the brand by name.

Attributed Branded Searches appear in the Demand Gen campaign report and are additive to, not a replacement for, view-through conversions and assisted conversion data. The metric is particularly useful for brand advertisers who have historically struggled to connect Demand Gen spend to a measurable business outcome in the Google Ads interface. However, because the attribution model relies on Google’s own identity graph and panel data, the figures should be treated as directional rather than deterministic, and should not be used as the sole basis for budget allocation decisions.

Advertisers outside the US should note that Attributed Branded Searches availability and data coverage may vary by market depending on Google’s search panel density in that country.

Channel Controls Now Broadly Available

Placement-level channel controls—allowing advertisers to include or exclude specific surfaces such as YouTube, Discover, Gmail, and Google Display Network—are now broadly available across Demand Gen campaigns.[7] This is a meaningful operational change from earlier in 2025, when placement controls were limited or absent, and Google’s AI managed the full allocation automatically.

Google’s default remains automatic allocation across all eligible surfaces, which the platform’s own guidance identifies as the highest-reach configuration.[10] However, channel controls give performance marketers a lever to pull when placement-level data shows a meaningful efficiency gap. For example, if the Gmail placement is consistently delivering CTR below 0.50%[9] while YouTube and Discover are at or above benchmark, an advertiser can now suppress Gmail without creating a separate campaign.

Worked example

Using Channel Controls to Suppress an Underperforming Gmail Placement Mid-Campaign

  • Setup: A San Francisco B2B SaaS account running a Demand Gen prospecting campaign at $6,000/month ($200/day) targeting US-based IT decision-makers aged 30–55. After 21 days of delivery (past the learning period), the placement breakdown report shows: YouTube CTR = 1.4%, Discover CTR = 1.9%, Gmail CTR = 0.28%. Overall campaign CPA for demo requests = $68.
  • Numbers: Gmail CTR of 0.28% is 44% below the 0.50% lower bound of the overall Demand Gen CTR benchmark range of 0.50%–2.00%.[9] Gmail’s share of total spend over 21 days = $420 (21% of $2,000 spent). At a Gmail CPC of $0.45–$0.95 (midpoint $0.70[5]), that $420 generated approximately 600 Gmail clicks—but zero attributed demo conversions in the same window, implying an effective Gmail CPA of undefined (division by zero). Reallocating the $420 Gmail spend to YouTube (CPC midpoint $0.875[5]) would generate approximately 480 YouTube clicks; at the campaign’s observed YouTube conversion rate implied by a $68 CPA and 1.4% CTR, this reallocation could generate an estimated 6–8 additional demo conversions over the same period.
  • Decision: Apply a channel control exclusion to Gmail effective immediately. Leave YouTube and Discover on automatic allocation. Re-evaluate placement mix after the next 14 days of data with Gmail excluded, checking whether blended CPA moves below $68.
  • Why: A Gmail CTR of 0.28%—44% below the 0.50% benchmark floor and producing zero conversions in 21 days—combined with 21% of budget consumed by that placement provides sufficient evidence to suppress the surface using the newly available channel controls without requiring a full campaign rebuild.[7][9]

Display-to-Demand Gen Migration: Final Push

Google has continued the formal migration of legacy Google Display Network campaigns into Demand Gen throughout 2026, and the August 2026 communications confirm that Demand Gen is now the primary visual advertising format within the Google Ads ecosystem.[7] Advertisers still running unconverted Display campaigns should treat migration as urgent, not optional. Google’s migration documentation provides a step-by-step upgrade path, and the Google Ads API release notes confirm that new API features are being built for Demand Gen rather than legacy Display objects.[12]

One structural implication of this migration that practitioners should understand: Demand Gen’s AI-driven delivery and creative assembly differ meaningfully from Display’s manual placement and creative controls. Accounts migrating from Display should expect a 2–4 week learning period after migration during which performance metrics may fluctuate, and they should avoid making bid or audience changes during that window per Google’s 14-day no-edit guidance for new bidding configurations.[3]

Worked example

Migration Budget and Learning Period Planning for a Legacy Display Account

  • Setup: A Houston home services franchise running a legacy Google Display campaign at $3,500/month ($117/day) with a manual CPC bid of $0.65 and a target CPA of $55 for appointment form submissions. The account has 38 Display conversions over the past 30 days. The franchise’s Google rep has flagged that the campaign type will be sunset and recommends migration to Demand Gen.
  • Numbers: 38 conversions in 30 days = 1.27 conversions/day—above the minimum threshold of ~2.1/day for stable tCPA (30 conversions/14 days) but within striking range if volume holds post-migration. 10× tCPA rule for the $55 target = $550/day recommended budget[1]; current budget of $117/day is 79% below that floor, meaning tCPA should not be the launch bidding strategy for the migrated campaign. Recommended launch strategy: Maximize Conversions at $117/day for 14 days, then evaluate whether conversion volume has reached 30 in 14 days before switching to tCPA $55.[3]
  • Decision: Migrate to Demand Gen on October 1, 2026. Launch with Maximize Conversions bidding at $117/day. Upload creative assets matching the 3-per-ratio requirement (vertical, square, landscape).[2] Make zero bid, audience, or budget changes from October 1–14. On October 15, pull the bid strategy report: if conversions ≥ 30 over the 14-day window, switch to tCPA $55; if conversions < 30, extend Maximize Conversions for another 14 days.
  • Why: A daily budget of $117 is 79% below the 10× tCPA threshold of $550/day; launching a tCPA campaign at this budget level virtually guarantees a permanent learning status—the same failure pattern identified in the preceding section—so Maximize Conversions is the correct launch strategy until volume and budget alignment can be achieved.[1][3]

What Practitioners Should Do Right Now

  • Audit vertical asset coverage immediately. With Multimodal Video Creation now generally available, there is no production-cost excuse for missing vertical (9:16) assets. Any ad group with fewer than 3 vertical video or image assets is leaving Shorts and vertical YouTube inventory on the table.[2][7]
  • Enable Attributed Branded Searches reporting. Add this metric to your standard Demand Gen reporting dashboard to begin building a baseline. Even if you do not use it for bidding decisions, the trend line will become a useful incrementality proxy over Q4 2026 and into 2027.[11]
  • Check placement performance at the channel level and use controls where justified. Now that channel controls are broadly available, a placement delivering CTR below 0.50%[9] and zero attributed conversions after 21 days should be suppressed rather than tolerated as a “reach” contribution.
  • If you are in travel, link your Hotel Center feed. Travel Feeds and personalized hotel ads are available now; accounts without a linked feed are missing dynamically priced creative that Google’s system can serve to high-intent travel audiences.[7]
  • Finalize any legacy Display migration before year-end 2026. Google’s direction is unambiguous: Demand Gen is the successor format. Delaying migration means operating on a campaign type that is receiving no new feature development and may face forced migration on Google’s timeline rather than yours.[7][12]
  • Monitor CTV impression share if you run shoppable Demand Gen. CTV CPMs price at the top of the YouTube range ($8.00).[5] If CTV share climbs above 20% of YouTube impressions without a corresponding budget increase, total reach will compress. Check the placement report weekly during the initial CTV ramp period.

References

  1. [1] https://blog.google/products/ads-commerce/demand-gen-drop-february-2026/ blog.google
  2. [2] https://support.google.com/google-ads/answer/14693848 support.google.com
  3. [3] https://support.google.com/google-ads/answer/14733311?hl=en support.google.com
  4. [4] https://roa-marketing.com/blog/google-ads-demand-gen-campaigns-2026/ roa-marketing.com
  5. [5] https://prospeo.io/s/demand-gen prospeo.io
  6. [6] https://blog.google/products-and-platforms/products/youtube/youtube-demand-gen-updates/ blog.google
  7. [7] https://blog.google/products/ads-commerce/demand-gen-drop-august-2026/ blog.google
  8. [8] https://support.google.com/google-ads/answer/13695777?hl=en support.google.com
  9. [9] https://almcorp.com/blog/google-demand-gen-best-practices-february-2026/ almcorp.com
  10. [10] https://www.astraloopstudio.com/en/blog/google-ads-demand-gen-guide/ www.astraloopstudio.com
  11. [11] https://blog.google/products/ads-commerce/demand-gen-drop-january-2026/ blog.google
  12. [12] https://developers.google.com/google-ads/api/docs/demand-gen/create-campaign developers.google.com
Share the Post:

Related Posts