Google Ads Demand Gen Shifts to CPM: What to Do

The Google Ads Demand Gen CPM billing change is the kind of update that hides in a settings tab and quietly rewrites your unit economics. On June 16, 2026, Google notified advertisers that Demand Gen campaigns using view-through conversion (VTC) optimization on the Discover placement will move from cost-per-click to cost-per-mille billing, effective July 15, 2026 — and it applies automatically, with no required action.

That single switch changes what you are paying for. Under CPC, you pay only when someone taps your ad. Under CPM, you pay for every thousand impressions whether or not anyone clicks. For a placement built on casual, low-intent feed browsing, that is not a cosmetic change — it is a different cost model that rewards different creative and punishes a different kind of waste.

This playbook covers exactly what changed, the narrow set of campaigns actually affected, the breakeven math that tells you whether CPM helps or hurts, the one setting that opts you out, and how to pressure-test whether view-through attribution is even worth paying impression rates for. Every figure here is sourced and dated — and where a benchmark is a proxy or a vendor claim, we say so.

On June 16, 2026, Google began notifying advertisers that Demand Gen campaigns using view-through conversion optimization on Discover will shift from CPC to CPM billing on July 15, 2026, with the change applying automatically and no required advertiser action. The news was reported by Search Engine Land, which is the source of record for the dates and scope below.

The change was first surfaced publicly by Anthony Higman, founder and CEO of ADSQUIRE, who shared Google’s advertiser communication on X. Google’s stated rationale, as reported, is that CPM billing “more accurately reflects the value being delivered” for view-through conversion campaigns — because view-through conversions are triggered by impressions seen, not by clicks taken. In other words: if you are optimizing toward conversions that happen after an impression rather than a click, Google argues you should be billed on the impression.

The timing is what makes this notable. VTC optimization for Demand Gen only launched as an open beta in Google’s April 2026 Demand Gen Drop — initially supporting YouTube inventory, with additional surfaces flagged to follow. Discover is that next surface. Advertisers who opted into the beta roughly eight weeks ago, when it was a YouTube-only bidding enhancement, are now seeing it carry a billing- model change on a different placement. That short runway is the real story for anyone managing these campaigns.

The most useful thing to internalize is how narrow the trigger is. Two conditions must both be true. First, the campaign must serve on the Discover placement. Second, view-through conversion optimization must be enabled. Miss either condition and nothing changes for you on July 15.

The second condition matters because VTC optimization is disabled by default for both new and existing Demand Gen campaigns. Per Google’s help documentation, only advertisers who explicitly turned it on after it launched in April 2026 are in scope. If you have never knowingly enabled view-through conversion optimization, the most likely answer is that you are not affected — but you should still verify in the campaign settings rather than assume.

One nuance worth flagging: Demand Gen is a multi-surface format. A single campaign can serve across YouTube, Gmail, Discover, the Google Display Network, and Maps, and you have limited granular control over exactly where impressions land. So “I do not really target Discover” is not the same as “my campaign never serves on Discover.” If VTC optimization is on and Discover is in your eligible inventory, treat yourself as potentially in scope and check the setting directly.

CPC and CPM are not two prices for the same thing — they are two different ways of assigning risk. Under cost-per-click, the platform carries the risk of a non-clicking impression: you only pay when someone acts. Under cost-per-mille, you carry it: you pay for the impression and hope it converts the audience downstream. For a view-through-optimized campaign, that reassignment is the whole point — VTC exists precisely to credit impressions that influenced a conversion without a click.

A view-through conversion is recorded when a user sees your ad impression without clicking, then converts within a defined lookback window. Google’s attribution hierarchy prioritizes clicks first, then engagements, then views — a view credit only applies if no click or engagement was recorded. The default VTC lookback window is one day, and you can set it between one and thirty days. Longer windows capture more conversions but risk crediting impressions that had limited causal influence on the decision.

Here is the practical consequence. Under CPC billing, high click-through creative effectively earns you a discount per impression — more clicks spread the cost-per-click bill across the value you capture. Under CPM, that relationship inverts: you pay the same per thousand impressions whether your CTR is 0.3% or 2.5%. The advertisers who lose most under CPM are the ones whose creative was quietly subsidizing them with strong click rates. The advertisers who may benefit are those whose impressions were genuinely driving view-through conversions that CPC billing never charged for. The next section turns that intuition into numbers.

No published source maps the Discover CPC-to-CPM transition into a breakeven, so we built one. The logic is simple. Under CPC billing, the amount you effectively pay per thousand impressions equals your cost-per-click multiplied by your click-through rate times one thousand. That is your effective CPM today. CPM billing is cheaper only when the rate Google charges per thousand impressions comes in below that number.

For the worked example below we use an illustrative Discover CPC of $0.50 — squarely inside the $0.20 to $0.80 ecommerce Discover CPC range that Store Growers reports. We then compare each row’s effective CPM against a proxy CPM band of $4 to $10. That band is not a Discover CPM benchmark — no Discover-specific CPM benchmark exists before the July 15 change. It is derived from YouTube formats (Shorts CPM around $4; non-skippable in-stream roughly $6 to $10) as the closest available proxy for what VTC-optimized Discover CPM rates might resemble. Treat the verdict column as directional, not a quote.

To see why the Discover change stands out, it helps to map the full picture. Demand Gen already uses different billing models by placement, and the July 15 change moves one specific cell. The table below pulls current billing from Google’s Demand Gen FAQ, the July 15 change from Search Engine Land, and the migration context from Search Engine Journal. No single published source maps all six placements alongside the forward trajectory — that connective view is the proprietary part.

Two things jump out. First, this is a single-cell change today — only VTC-optimized Discover image and video billing moves on July 15. Second, the direction of travel is unmistakable: every other Demand Gen video surface already bills on CPM, and Display, which has always been impression-capable, is being absorbed into Demand Gen. The Discover shift reads less like an isolated tweak and more like Google standardizing toward impression-based billing across the format. We cover that consolidation in Section 08.

If you decide CPM billing does not fit your campaign, the opt-out is straightforward: disable view-through conversion optimization. According to Google’s help documentation, you navigate to the campaign settings, open the conversions-optimization controls, uncheck the option to include view-through conversions, and save. Disabling VTC optimization prevents the CPM billing change from taking effect on that campaign, because the change is tied to VTC optimization being on.

Two operational caveats are worth knowing before you lean on view-through conversions either way. VTC optimization on Demand Gen does not currently support offline conversions, store visits, or OCI conversion types, and it does not affect Conversion Lift studies. And browser tracking restrictions cut into VTC measurement: Safari blocks cross-site cookies, which reduces view-through visibility for a meaningful slice of traffic — a real concern if your audience skews mobile and iOS-heavy. If a large share of your view-through conversions are invisible to begin with, paying impression rates to optimize toward them deserves extra scrutiny.

The deciding question is not really “CPC or CPM” — it is whether your view-through conversions represent genuine causal influence or just incidental impressions next to organic conversions. If VTC is measuring real lift, CPM billing aligns cost with value. If VTC is mostly crediting impressions that would have converted anyway, you are about to pay impression rates to optimize toward a vanity number. So validate the signal before the billing model locks in.

Industry data suggests view-through conversions are a material share of measured performance. AdNabu cites roughly 25% of total conversions in Google Display and Video campaigns being attributed to view-through — a vendor-cited figure, so treat it as directional rather than independently confirmed. Google, for its part, has stated that VTC campaigns can add “another 20 to 40 percent” to campaign impact; that is a vendor claim, and practitioners widely recommend discounting it heavily for attribution-causality concerns. The number you should trust is the one you measure yourself.

This is also the cleanest place to bring in incrementality as a discipline rather than a one-off. The whole premise of view-through billing is that impressions cause conversions — a causal claim, not a correlational one. The only honest way to test a causal claim is a holdout. Our deeper treatment of incrementality testing to separate view-through attribution from organic lift walks through the holdout design that turns “VTC says it worked” into “the test proved it worked.” If you are about to start paying impression rates for view-through conversions, that proof is worth having first.

Step back and the Discover CPM shift looks less like an isolated billing tweak and more like a leading indicator. Google is migrating standalone Display campaigns into Demand Gen, with a voluntary in-platform migration tool available from June 2026 and a manual migration deadline of January 2027; automatic migration of remaining campaigns continues through 2027. Display has always been impression-capable. As Demand Gen absorbs it, impression-based billing becomes the natural common denominator across the format’s surfaces.

That consolidation is also why the format itself keeps growing in advertiser importance. Demand Gen can reach up to three billion monthly active users across YouTube, Gmail, Discover, the Display Network, and Maps. Google-cited and partner data point to it pulling distinct, incremental audiences — for example, vendor figures suggest a large share of Demand Gen conversions come from users who had not seen the brand’s Search ads recently, and a Google-commissioned Fospha study reports an 18% higher share of new-customer conversions versus the paid-media average. Those are vendor-stated and vendor-commissioned numbers, so we cite them as directional context, not as independent proof — but the strategic point holds: this is a format Google is investing in and standardizing, not winding down.

Here is our forward read. The Discover CPM change is unlikely to be the last impression-based billing move in Demand Gen. As Google unifies Display and Demand Gen and leans harder into view-through and engaged-view optimization, expect more of the format to bill on attention rather than clicks over the next year. For advertisers, the durable response is not to fight each individual change but to rebuild measurement around incrementality, so that whichever billing model applies, you already know what a thousand impressions are actually worth to your pipeline. If you want help re-modelling budgets under shifting platform rules, that is core to our paid media management work, and it pairs naturally with the measurement rigour in our analytics and measurement engagements.

For the wider budgeting context, two companion pieces are useful here: our channel-level budget allocation framework for re-slicing spend when a channel’s cost model changes, and the in-house vs. retail media budget allocation matrix for deciding where Demand Gen sits in the overall mix. If you’re building the upper-funnel case at all, start from the mid-funnel demand generation benchmarks.