Google Demand Gen Attribution: Measuring a Campaign Built to Take Credit
Demand Gen reports conversions across YouTube, Shorts, Discover and Gmail using view-through and data-driven attribution inside Google's walls. Here is why those numbers inflate, the four layers where credit leaks in, and how to measure the channel's real incremental sales.
Three platforms, three answers, one budget to defend. Your GA4 export already holds the answer, and the first finding is free and never leaves your browser. Get your first finding free
By Joris van Huët, Founder & CEOUpdated 6 min read
Run the numbers for your store: the free view-through attribution calculator.
The 60-Second Answer
Google Demand Gen attribution credits conversions across YouTube, Shorts, Discover and Gmail using view-through conversions and data-driven attribution measured only inside Google's ecosystem. Because it counts post-impression conversions and cannot see non-Google touchpoints, its reported ROAS systematically overstates the campaign's true incremental contribution. Validating with a holdout or causal analysis is the only reliable fix.
What Demand Gen Is: and Why Its Numbers Are Different
Demand Gen is the Google Ads campaign type that fully replaced Discovery campaigns in 2024. It serves visually-driven ads across YouTube (including Shorts), Discover, and Gmail: Google describes reach of more than 3 billion monthly users and over 50 billion daily Shorts views. It is, in effect, Google's answer to paid social: an upper- and mid-funnel demand-creation channel that lives inside a walled garden.
That positioning is exactly why its attribution behaves differently from search. Search captures existing intent and is mostly measured by clicks. Demand Gen creates interest, so Google measures it heavily through impressions and views: which means view-through conversions and data-driven attribution do most of the credit-assigning. Both mechanisms favor the platform doing the reporting.
In April 2026 Google added view-through conversion optimization for Demand Gen on YouTube, letting bidding optimize toward conversions that happen after an ad impression with no click at all. Google also introduced an "attributed branded searches" metric (rolled out from January 2026) that credits Demand Gen with the branded searches it claims to generate. Useful signals: but each one widens the gap between reported and incremental.
The Demand Gen Attribution Gap Stack (Original Framework)
| Layer | What it does | Direction of distortion |
|---|---|---|
| 1. View-through layer | Counts post-view conversions from un-clicked impressions | Inflates: many buyers would have converted anyway |
| 2. Cross-device layer | Stitches logged-in Google identities you cannot independently verify | Inflates: claims journeys your own tools can't see |
| 3. In-walls DDA layer | Data-driven attribution models credit using only Google touchpoints | Inflates: Meta, email and organic are invisible to it |
| 4. Branded-search layer | "Attributed branded searches" credit demand that may already exist | Inflates: harvesting is counted as creation |
The deeper you go, the more the stack converts correlation with a purchase into claimed responsibility for it. No single layer is fraudulent; together they routinely double a channel's apparent return. This is the correlation-versus-causation problem wearing a new campaign label.
How Demand Gen Reports vs What Actually Happened
| Placement | How Google credits it | What it usually misses |
|---|---|---|
| YouTube in-stream | View-through + clicks | Whether the viewer already intended to buy |
| Shorts | View-through (short, cheap impressions) | large impression volume inflates VTC |
| Discover feed | Click + engaged-view | Cross-channel assists from Meta and email |
| Gmail | Open/click | Existing customers re-converting |
Compare this to Performance Max, which shares the same in-walls measurement DNA, and to YouTube view-through attribution more broadly. The pattern is identical: the platform sees its own impression, sees a later conversion, and connects them: even when your MTA tool disagrees.
How to Measure Demand Gen's Real Incremental Value (Step by Step)
- Isolate the channel. Use Google's Campaign Type Attribution to separate Demand Gen conversions from Search and PMax, so you are not comparing blended numbers.
- Split click-through from view-through. Report post-click and post-view conversions separately. Never let a single ROAS figure hide the mix.
- Match the window to your buying cycle. A 30-day view-through window on a 5-day purchase cycle invents assists. Align it with your real attribution window.
- Run a geo or audience holdout. Suppress Demand Gen in matched regions and measure the delta in total orders: the logic of a Meta-style holdout applies directly. This is the cleanest incrementality test.
- Cross-check against a portfolio metric. If Demand Gen's reported revenue rises but your MER and new-customer CAC do not improve, the credit is not incremental.
- Run causal attribution on your GA4 export. Estimate Demand Gen's contribution against a modeled counterfactual instead of trusting in-walls data-driven attribution.
A €-Denominated Worked Example
A Shopify apparel brand runs Demand Gen at €10,000/month. Google Ads reports:
| Metric | Google Ads (in-walls) |
|---|---|
| Reported conversions | 500 |
| Reported revenue | €60,000 |
| Reported ROAS | 6.0x |
| Share from view-through | 58% |
A 6x ROAS looks like a clear "scale it." But 58% of the revenue is view-through, and the brand also runs heavy Meta prospecting that Google's model cannot see.
The team runs a four-week geo holdout: Demand Gen paused in matched regions. Total orders fall by far less than Google's numbers imply.
| Metric | Google reports | Holdout-measured truth |
|---|---|---|
| Revenue credited to Demand Gen | €60,000 | €23,000 incremental |
| True ROAS | 6.0x | 2.3x |
| Decision | "Scale aggressively" | "Profitable but modest: scale carefully" |
The channel is still worth running: but at 2.3x, not 6.0x. Doubling the budget on the reported number would have pushed real ROAS below break-even while the dashboard kept flashing green. (Figures are illustrative.)
Common Mistakes
- Reading reported ROAS as incremental ROAS. Demand Gen's headline number is the top of the Gap Stack, not the bottom.
- Comparing Demand Gen and Meta on each platform's own numbers. Both over-claim; the comparison is meaningless without a neutral referee.
- Leaving long view-through windows on by default, inflating credit for short-cycle purchases.
- Treating "attributed branded searches" as net-new demand when much of it is harvested.
- Optimizing bids toward view-through conversions without ever validating that those conversions were incremental.
- Ignoring cross-device double-counting between YouTube and other Google surfaces.
Checklist: Auditing Your Demand Gen Numbers
- Demand Gen isolated via Campaign Type Attribution
- Click-through and view-through conversions reported separately
- View-through window aligned to your real purchase cycle
- At least one geo or audience holdout on the books
- Reported revenue sanity-checked against MER
- Cross-validated with causal attribution on your GA4 export
- Demand Gen compared to other channels on one neutral cross-channel view
Key Takeaways
- Demand Gen is a demand-creation channel measured largely by impressions, so view-through conversions and in-walls data-driven attribution drive most of its credit.
- Its reported ROAS sits at the top of a four-layer Attribution Gap Stack and systematically overstates incremental value.
- New 2026 features, view-through optimization and attributed branded searches, improve activation but widen the reporting-versus-reality gap.
- The only reliable validation is a holdout or causal attribution read against a counterfactual: the same lesson behind Meta Conversion Lift and MMM.
- Treat Demand Gen like any other walled garden: useful for activation, untrustworthy for self-scored measurement. See also Markov chain attribution and the unified measurement framework.
For €99, upload any historical GA4 period and get causal attribution for every channel in 1–2 minutes: no pixel, no migration.
Frequently asked questions
What is Google Demand Gen attribution?
It is how Google Ads credits conversions to Demand Gen campaigns running across YouTube, Shorts, Discover and Gmail. Because Demand Gen is an impression-led demand-creation channel, Google relies heavily on view-through conversions and data-driven attribution measured inside its own ecosystem, which tends to overstate the campaign's incremental contribution.Why is Demand Gen ROAS higher in Google Ads than in my other tools?
Google counts post-impression view-through conversions and stitches cross-device journeys it can see but you cannot, and its data-driven model only includes Google touchpoints. Your GA4 or MTA tool sees the full journey, including Meta and email, so it assigns Demand Gen less credit. The gap is structural, not a bug.Did Demand Gen replace Discovery campaigns?
Yes. Google migrated all Discovery campaigns to Demand Gen during 2024. Demand Gen expanded the available inventory to include YouTube in-stream and Shorts alongside the original Discover and Gmail placements, making it Google's primary visual, social-style demand-creation campaign type.What is view-through conversion optimization for Demand Gen?
Announced for YouTube in April 2026, it lets Demand Gen bidding optimize toward conversions that occur after an ad impression even when there was no click. It can increase reported conversions, but those view-through conversions are exactly the ones most likely to have happened anyway, so they should be validated with a holdout.How do I measure the true incremental value of Demand Gen?
Run a geo or audience holdout: pause Demand Gen in matched regions and measure the change in total orders. Report click-through and view-through conversions separately, align the attribution window to your real buying cycle, and cross-check against MER and new-customer CAC. Causal attribution on your GA4 export estimates contribution against a counterfactual.Are attributed branded searches a reliable Demand Gen metric?
Attributed branded searches, introduced from January 2026, show how many branded searches Google believes a Demand Gen campaign generated. It is a useful directional signal of interest lift, but some of those searches reflect demand that already existed, so treat it as evidence to test rather than proof of net-new demand.Is Demand Gen worth running if its reported ROAS is inflated?
Often yes. Inflated reporting does not mean zero value; it means the headline number overstates incrementality. Many brands find Demand Gen profitable at a lower true ROAS than Google reports. The goal is to size the budget against measured incremental return, not the dashboard figure.
Go deeper: Causal attribution, explained.
Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.
Have an audience of ecommerce operators? Earn 20% recurring commission by recommending the €99 read.
Keep reading
Terms in this article
- Ad ImpressionAd Impression is a single instance of an advertisement displaying on a webpage. Impressions are a key input for models measuring the causal impact of ad exposure on user behavior.
- AttributionAttribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
- Attribution ModelAn Attribution Model defines how credit for conversions is assigned to marketing touchpoints. It dictates how marketing channels receive credit for sales.
- Attribution WindowAttribution Window is the defined period after a user interacts with a marketing touchpoint, during which a conversion can be credited to that ad. It sets the timeframe for assigning conversion credit.
- Causal AnalysisCausal Analysis identifies true cause-and-effect relationships in data, moving beyond correlation to show how marketing actions directly impact outcomes.
- Causal AttributionCausal Attribution uses causal inference to determine which marketing touchpoints genuinely cause conversions, not just correlate with them.
- CounterfactualA counterfactual is what would have happened under conditions that did not occur, such as a customer's outcome had they not received the campaign they did receive. It is never observed, so every effect estimate has to stand in for it with a control group, comparable units or a forecast.
- IncrementalityIncrementality is the part of an outcome, such as orders or revenue, that a marketing activity caused: the extra results compared with what would have happened without it. It is a causal effect, not the number of conversions credited to a channel.