Scaling YouTube Campaigns in DV360: Best Practices for 2026 (Part 3)

AI Features in DV360: Automating Campaign Management

Within the DV360 interface, AI-powered capabilities include:

• Generate report: Skip manual filter setup. Describe the report you need in natural language, and DV360 will generate the relevant reporting view across Line Items or YouTube Ad Groups.

• Audience Persona: Describe your ideal audience, and AI will instantly build a combination of audience segments aligned with your goals. It automatically selects and combines relevant audience lists and demographic criteria, so you don’t have to manually browse hundreds of segments. This helps advertisers quickly identify the most relevant audiences from thousands of options and uncover new high-potential segments.

• AI-powered search: Replaces traditional filters with semantic search. It lets you find relevant campaign entities based on settings, targeting, reporting data, and event history. Simply type a query such as “paused line items” to surface the results you need.

• AI Chat Assistant: A virtual assistant built directly into the workspace that answers technical questions in real time, explains how the platform works, and helps with YouTube campaign setup.

• Targeting Templates:Create, edit, and save templates at the Advertiser level for use across multiple campaigns. You can create new Line Items from templates or apply a template in bulk to existing Line Items.

As a result, AI in DV360 is gradually taking on part of the operational work — from finding the right settings and building audiences to preparing analytics.

Practical Case Study: Gen AI Advertising for Lay’s

Let’s look at a real-world example of a major brand running a YouTube campaign through Display & Video 360.

Lay’s set out to test whether Gen AI creative could become an effective complement to traditional video production while also exploring its potential to reduce production costs. To do this, Lay’s, together with VIVID and CoobX, developed an “AI + Precision” strategy that combined generative AI-powered video production with precise media planning.

Using Google Veo 3, the teams created a 30-second video ad titled “Battle of Flavors.” To distribute the creative across the Google Marketing Platform ecosystem, the team used a differentiated enterprise approach, splitting the campaign into two activation tracks:

• For reach: a CPM model using YouTube Efficient Reach.
• For views: a CPV model through Video Views.

The activation results demonstrated both the cost efficiency and strong media performance of the Gen AI creative. Optimizing toward views (CPV) enabled the brand to generate video views at an exceptionally low cost, with eCPV of just $0.002.

More importantly, cost optimization and AI-powered production did not come at the expense of audience engagement, as reflected in the key performance metrics:

• Completion Rate reached an impressive 65%;
• TrueView Rate reached 64.7%;
• Average watch time reached 24 seconds for a 30-second ad, indicating strong audience engagement with the creative.

The Lay’s case demonstrated that Gen AI video can effectively complement the traditional media mix, reduce content production time, and still deliver strong engagement. 

YouTube Campaign Measurement and Optimization in DV360

Following YouTube campaign best practices is only part of the equation — measuring performance correctly is just as important. Since DV360 supports the full marketing funnel, the measurement framework can be divided into four core areas:

• Brand Lift & Search Lift: measuring brand awareness and brand impact.
• Video Metrics: evaluating engagement quality and attention retention.
• Unique Reach & Frequency: measuring unique reach and exposure frequency.
• Conversions & Attribution: tracking lower-funnel performance outcomes.

Let’s look at each area in more detail.

Brand Lift

For upper- and mid-funnel campaigns, standard performance metrics are not always enough to understand brand awareness or the associations audiences have with a brand.

Brand Lift studies automatically divide the audience into two groups: users who saw the ad (exposed) and users who were eligible for targeting but did not see the ad (control). Based on YouTube user surveys, the system measures six key metrics:

• Ad Recall: whether users remember seeing your ad.
• Awareness: how familiar consumers are with the brand or product.
• Consideration: whether consumers would consider the brand as an option for a future purchase.
• Favorability: whether consumers have a positive perception of and emotional connection with the brand.
• Purchase Intent: the likelihood that consumers will purchase the brand’s product in the near future.
• Association Lift: a new mid-funnel measurement tool introduced in H1 2026 that evaluates which values, ideas, or attributes consumers associate with the brand.

Search Lift
The mechanics are straightforward: advertisers define a list of keywords and launch the campaign, while Search Lift measures how much more frequently those terms are searched by users who saw the ad compared with those who did not.

Video Metrics: Measuring Engagement Quality

To assess how well the creative holds viewers’ attention, we use a broader set of video metrics. The primary indicators include TrueView views and VTR (View Rate):

VTR = Views / Impressions

Additional metrics for evaluating video perfomance include:

• Video Played to 25/50/75/100% (Quartile reporting): clearly shows where viewers drop off throughout the video. A sharp decline within the first 5 seconds may indicate a weak hook, while a drop after the 25% mark may point to weak narrative development in the middle of the video.
• Avg. Watch Time per Impression: the average number of seconds watched per impression. A steady decline in this metric is a key indicator of Creative Fatigue.
• Earned Actions: actions viewers take after watching the ad, including likes, shares, subscriptions, and playlist additions. This metric helps indicate the viral potential and organic value of your content.
• YouTube Public Views: the public view count displayed on the video. Following a platform update, Bumper and Non-skippable formats can now also contribute to YouTube Public Views, although they do not generate TrueView Views. Skippable formats can contribute to both metrics.

Dimensions & Splits: Common Measurement Pitfalls

For accurate analysis, advertisers should avoid looking at a single blended VTR across all formats. In-stream counts a view after 30 seconds, while In-feed and Shorts count views after 10 seconds and follow different measurement logic. Combining these metrics can therefore be misleading. For example, within a single Line Item, In-stream may deliver a VTR of 60%, while In-feed delivers 10%. Looking only at the overall VTR for the Line Item would produce an average that does not accurately reflect the performance of either format. VTR should therefore always be analyzed across the following Dimensions:

• YouTube Ad Format — breakdown by format: in-stream, in-feed, Shorts, bumper;
• Device Type — certain formats and video creatives consistently perform better on specific devices;
• Ad Group — for comparing ad groups and creatives within a single Line Item;
• YouTube Video — for evaluating the performance of individual video assets;
• Media Type — breakdown by Audio and Video media types.

Also Keep the Following Platform-Specific Measurement Considerations in Mind:

• QR codes on CTV = clicks: Since Connected TV does not support traditional clicks, when a viewer scans a QR code displayed on the TV screen with a smartphone, the system records the interaction as a standard click in reporting.
• Quartiles ≠ Views: Do not calculate View Rate from quartile data, as these metrics are calculated using different methodologies. For In-feed, “video played to” metrics are calculated from the moment the video player starts rather than from impressions.
• Invalid views (adjustments for up to 30 days): YouTube applies strict anti-fraud filters and may adjust or remove invalid views related to invalid traffic for up to 30 days after delivery. If metrics decrease retrospectively, this may be part of the platform’s standard verification process designed to protect advertiser budgets.

Reach & Frequency and the Target Frequency Tool

Unique reach is evaluated through reach metrics. In media planning, gross Impressions without the context of Reach provide limited value: 1,000,000 impressions could represent 100,000 users exposed 10 times each, creating overexposure, or 500,000 unique users exposed twice each, delivering five times the unique audience reach.

A major advantage of DV360 reporting is its ability to deduplicate audience reach across campaigns, channels, and devices: one person = one reach, regardless of where they were exposed to the ad.

Key reach metrics include:

• Unique Users Reached — the number of unique deduplicated users reached;
• Avg Impression Freq / User (7d) — average weekly impression frequency per user;
• Cost Per Unique User Reached — cost per unique user reached;
• Unique Reach % of Target — the percentage of the target audience reached;
• Incremental Reach — additional reach delivered by YouTube on top of other channels.

When running campaigns on large screens, advertisers should also account for co-viewing, as multiple people often watch the same CTV device simultaneously. As a result, the traditional assumption of one device = one person no longer applies.

According to research from Analytic Partners + Google Meridian, the optimal weekly frequency benchmark for most brands ranges from 2.7 to 3.5 impressions per week. More than half of advertisers operate below this range, even though increasing frequency by 40% can improve overall ROI.

To manage frequency more precisely, use a Target Frequency line item. Unlike a standard frequency cap, which simply limits excess impressions, Target Frequency uses AI to optimize delivery toward the average frequency defined in the media plan, helping advertisers reach users the desired number of times over a given period. During peak season, optimal frequency may be 2× higher than usual, so frequency planning should account for seasonality rather than remain static.

Lower Funnel: Floodlight for YouTube

To ensure accurate conversion tracking and enable algorithmic optimization, advertisers need to activate Floodlight data collection. In DV360 Advertiser Settings, navigate to:

Advertiser Settings → Basic Details → YouTube Settings and enable: Use Floodlight data for conversion optimization on Demand Gen line items.

Without this setting, conversions will still appear in reporting, but they cannot be selected for bid optimization. The platform distinguishes between three types of YouTube conversions:

• Click-through: counted when a user clicks the ad and completes the desired action on the website within the defined conversion window.
• Engaged View (EVC): a non-click conversion. This attribution type is widely used for video campaigns: the user does not click the ad immediately but watches it for a sufficient amount of time and then converts within the conversion window. For Skippable In-stream, the engagement threshold is 10 seconds (or the full video), while for In-feed and Shorts it is 5 seconds. For Bumper and Non-skippable formats, EVC is available only when a click occurs. This is a critical metric for video advertising, as users rarely make a purchase immediately after seeing an ad and often convert several days later.
• View-through: records an Impression without direct interaction. Currently, these conversions are available in reporting but cannot be used for bid optimization.

Keep Testing: New Formats, Features, and Approaches

Finally, building a sustainable competitive advantage requires an ongoing cycle of experimentation. Testing new formats and platform features can give brands a twofold advantage:

• Lower competition at launch: when a new format is first released, fewer advertisers are using it. This can reduce auction CPMs and allow brands to access more high-quality inventory within the same budget.
• Novelty effect: users may pay more attention to new advertising formats, which can positively influence engagement and create additional potential for ROI growth.

YouTube within Display & Video 360 provides advertisers with more advanced capabilities for tracking and analyzing video campaigns, but using them effectively requires strong platform expertise and a clear understanding of the underlying metrics.

If you want to improve your media split, strengthen conversion attribution, explore new video advertising formats, or optimize video campaigns for higher ROAS, the CoobX team is ready to be your trusted technology partner.

Contact us.