Google Demand Gen 2026: Messaging, Creative and VTC Measurement for Indian D2C

Where can Demand Gen campaigns appear now?

Google's current Demand Gen campaign documentation lists YouTube, including Shorts, Discover, Gmail, Maps and the Google Display Network. Advertisers can choose channels and use horizontal, portrait, square and vertical creative, carousel ads and product feeds where supported.

That breadth creates a common reporting mistake: treating Demand Gen as one homogeneous placement. A vertical creator video on Shorts, a product card on Discover and a responsive image on the Display Network do not have the same user experience or attribution pattern.

Build the reporting view around four dimensions:

| Dimension | Minimum split | Decision it supports | | --- | --- | --- | | Channel | YouTube, Discover, Gmail, Display and other eligible surfaces | Where delivery and outcomes differ | | Format | Shorts, in-feed, in-stream, image, carousel and product feed | Which creative system earns attention | | Event type | Click, engaged view and impression | How Google assigned conversion credit | | Commercial outcome | Gross orders, net orders, new customers and contribution | Whether the campaign fits the business |

How should Indian D2C teams test messaging from Demand Gen?

If the account exposes the messaging option, define the test as a lead-to-order workflow rather than a click experiment.

1. Confirm the destination. Record the messaging app, account, supported geography and business-hours coverage.
2. Define a qualified conversation. Separate accidental opens and generic questions from conversations that include a product, use case, location or purchase constraint.
3. Preserve source data. Pass a stable campaign and creative identifier into the conversation or CRM record where the platform permits it.
4. Set a response-time rule. A media test cannot be evaluated fairly if high-intent conversations wait hours for a reply.
5. Reconcile orders. Match qualified conversations to net orders using a disclosed window and deduplication rule.
6. Keep a comparison group. Compare against a campaign, geography, audience or time cell without messaging when scale permits.

Do not call a conversation incremental merely because it began from an ad. The user may already have intended to buy, may have seen other campaigns, or may have converted through another channel.

How should multi-format creative be measured?

Google's asset-reporting guide warns that asset-level clicks should not be summed as if each represented a different ad click. A responsive ad can serve several assets together, causing the same click to appear against multiple components. It also says product-feed campaigns may attribute performance to product images rather than ad-level assets.

Use asset reporting for diagnosis, not finance reconciliation:

• compare assets within the same asset type;
• segment campaign results by format and channel;
• track spend, impressions, qualified views, clicks and attributed value;
• rotate a limited number of hypotheses at a time;
• log every creative introduction and removal date; and
• use a clean experiment when the decision is large enough to justify one.

The creative-fatigue measurement guide provides a rotation framework. The short-form video strategy explains how to build distinct hooks and formats rather than resizing one master asset mechanically.

What does view-through conversion optimization change?

Google's view-through optimization documentation says the feature is in open beta and available to interested advertisers for supported YouTube, Display and Discover inventory. New Demand Gen campaigns have view-through optimization enabled by default for video assets.

The attribution hierarchy is:

1. click-through conversion;
2. engaged-view conversion; and
3. view-through conversion.

Google describes a view-through conversion as a conversion matched to an impression within the selected window when no click or qualifying engagement takes precedence. The Help Center recommends reviewing the window and notes that a longer window can increase the number reported. It also warns that impression-based outcomes may not be exported into a third-party backend, creating a discrepancy between Google Ads and the order system.

Most importantly, Google's documentation explicitly says view-through optimization does not directly focus on incrementality. That boundary should appear in every performance review.

What is a defensible Demand Gen scorecard?

Use three layers rather than one blended ROAS number:

| Layer | Measures | Claim allowed | | --- | --- | --- | | Platform delivery | Spend, reach, views, clicks, CTC, EVC, VTC and attributed value | How Google delivered and assigned credit | | Business truth | Net orders, new customers, contribution and returns | What the business recorded under its rules | | Causal evidence | Holdout lift, conversion lift or another valid counterfactual | What likely happened because of the campaign |

The attribution glossary explains why assigned credit is not the same as cause. Use the Incrementality Lift Calculator to document exposed and control outcomes when a valid comparison exists.

What does a worked example look like?

Consider a fictional Indian skincare brand running a four-week Demand Gen test with video, image and product-feed assets. The figures below are synthetic and are not a benchmark.

| Measure | Platform report | Backend or test evidence | | --- | ---: | ---: | | Spend | ₹6,00,000 | ₹6,00,000 reconciled | | Click-through conversions | 210 | Included in deduplicated orders where matched | | Engaged-view conversions | 95 | Not independently exported to the order ledger | | View-through conversions | 180 | Not independently exported to the order ledger | | Platform-attributed conversions | 485 | Attribution total, not unique incremental orders | | Net matched orders | — | 330 after cancellations and duplicates | | Exposed-region net orders | — | 2,430 | | Matched-control expected orders | — | 2,280 | | Estimated incremental orders | — | 150 before significance and contamination checks |

Method: freeze the primary conversion definition and view-through window before launch; tag channel, format and creative; reconcile transaction IDs and cancellations; construct a matched control using prior demand and commercial conditions; exclude regions with major stock or promotion differences; and report uncertainty instead of presenting the 150-order difference as exact.

The example does not prove that view-through conversions are invalid. It shows why the 485 platform-attributed conversions, 330 net matched orders and 150 estimated incremental orders answer different questions.

How should product feeds be prepared?

Google's Demand Gen product-feed guidance recommends accurate Merchant Center data, eligible products for the targeted country and sufficient relevant assortment. It says campaigns can technically run with one product but recommends at least four for broader placement eligibility. Square product images are recommended for coverage.

Before launch, verify:

• price, availability, currency and landing-page consistency;
• product identity and variant grouping;
• country eligibility and shipping settings;
• image quality across square and other required ratios;
• exclusions for low-stock or low-contribution products; and
• stable product-level reporting identifiers.

HML's Merchant Center AI Performance Insights guide covers feed-quality diagnosis beyond paid delivery. The performance-marketing furniture case study provides context on aligning campaign structure with commercial outcomes; its results are not evidence for the synthetic example above.

What should the account owner do this week?

Create a one-page Demand Gen control sheet:

• feature availability and geography;
• channel and format plan;
• messaging qualification and response-time rules;
• creative hypothesis and rotation dates;
• view-through setting and conversion window;
• product-feed eligibility and margin exclusions;
• platform, backend and incremental reporting rows; and
• named owner for every discrepancy.

If Google Ads conversions cannot be reconciled with net orders or the team cannot distinguish click, engaged-view and view-through outcomes, request a Demand Gen measurement audit from HML's performance marketing team. The deliverable should be a traceable scorecard and test plan—not another dashboard that repeats the platform total.

Sources

1. Reach more customers with August's Demand Gen Drop — Google Ads & Commerce Blog; published 27 August 2026; accessed 23 September 2026.
2. About Demand Gen campaigns — Google Ads Help; accessed 23 September 2026.
3. About asset reporting for Demand Gen campaigns — Google Ads Help; accessed 23 September 2026.
4. About view-through conversion optimized bidding for Demand Gen campaigns — Google Ads Help; accessed 23 September 2026.
5. Use a product feed to show your products in Demand Gen campaigns — Google Ads Help; accessed 23 September 2026.