Google Demand Gen Gives You No Lead Form and No Call Button. Here Is How to Run It Anyway.

Google Demand Gen is Google’s campaign type for buying visual, interest-based inventory. Google’s own description is short: “Demand Gen campaigns serve ads on YouTube, Discover, Gmail, and the Google Display Network.” (About Demand Gen campaigns)

As of 2026 it is also the only way to buy that inventory. Google has folded Display into it and finished folding Video Action Campaigns into it.

Here is the part every other guide skips. Demand Gen hands a lead buyer almost no capture mechanism. There is no lead form asset. There is no call button. Sitelinks are the only extension. The one exception runs on a television. Google now generates a QR code and a “send to phone” button on Demand Gen ads. But those codes “are exclusively available for Connected TV (CTV)” and need TV screens in your device targeting. Google says they are “optimized to appear only when your ad is viewed on a TV” (QR codes for Connected TV). Send to phone transmits a link, not a dial action. That is a response captured from a TV screen, not a phone call. On the phone in a Shorts viewer’s hand you still get nothing.

So every lead and every call arrives on your own landing page. Google counts neither one unless you import it. And in the regulated verticals, the targeting rules quietly collapse the whole thing into a YouTube-only campaign.

That single constraint reorganizes everything else: your budget, your bidding, your creative, your measurement, and whether the channel is worth running in your vertical at all. This guide is built around it.

Quick answers:

• What are Google Ads Demand Gen campaigns?
• How is Demand Gen different from Performance Max?
• Does Google Demand Gen have lead form assets?
• Can you run call ads in Demand Gen?
• How much budget does a Demand Gen campaign need?
• Does Demand Gen work for lead generation?

What Demand Gen is now, and why you are being moved into it

Demand Gen started as the replacement for Discovery campaigns. It has since become the container Google is pouring its other visual campaign types into.

Display went first. Google’s wording is blunt: “Google Display Ads campaigns have a new home as Google Display Network (GDN) in Demand Gen.” The same page says “New Google Display Ads campaigns can only be created within Demand Gen.” (Display campaigns move to Demand Gen)

Three details in that migration matter more than the headline.

It is one-way. Google states plainly that “Campaigns that are migrated to Demand Gen can’t be reverted back to Google Display Ads campaigns.”

You inherit GDN whether you want it or not. “Migrated Google Display Ads campaigns will have Google Display Network opted in by default, and you will not be able to unselect it during migration.”

Your history comes with you. The migration tool ports performance history “dating back to 42 days,” which Google says “minimizes learning time to approximately 1-2 days and avoids a ‘cold start.'”

Video Action Campaigns are already done. Google confirms it is “automatically upgrading the final campaigns to Demand Gen by April 2026.” (Video Action Campaigns upgrade)

So if you run lead-gen on Google’s visual surfaces, you are going to end up here. The question is not whether to adopt Demand Gen. It is what you rebuild before you land in it.

We covered the migration mechanics step by step in the Display to Demand Gen migration playbook. What specifically breaks for lead buyers is in what the migration does to a lead-buying account.

Where your ads actually run, and how little you control it

Demand Gen channel controls sit at the ad group level, not the campaign level. (Channel controls) You either accept “All Google channels” or pick from YouTube in-stream, YouTube in-feed, YouTube Shorts, Discover, Gmail, Maps and the Google Display Network, though Google’s own Shorts buying page says you cannot buy Shorts on its own (see the contradictions below).

Two conditions attach to that list.

Maps is conditional. Google says Maps “only serves when a non-affiliated location extension is enabled on the account.” For a home services advertiser this is the single most call-relevant surface in the campaign, and it is off unless you have set up location assets.

Google steers you away from choosing. The same page calls automatic channel selection “the recommendation for most advertisers.”

There is a real cost to overriding that. Demand Gen uses marginal cost optimization, which chases the cheapest available conversion across surfaces.

Google’s own remedy is telling. To force volume onto a specific channel, “isolate the channel into a separate ad group with an increased CPA compared to the others.” You do not get a channel bid modifier. You get a second ad group and a higher price.

What you do not get anywhere is granular safety control. Demand Gen has no frequency capping at all. Asked whether you can control impression frequency, Google answers “No. The system automatically controls impression frequency.” (Demand Gen FAQ)

You also cannot set location and language independently: “You can only set the location or language at the campaign or ad group level (never both).”

Decision rule: start on All Google channels to let the system find the pocket of demand, then isolate a channel into its own ad group only once your own backend data, not Google’s, shows that surface producing qualified units. If you are already seeing junk placements, run the Demand Gen placement exclusion audit before you spend another month.

The thing every other guide leaves out: no lead form, no call button

This is the constraint that defines the channel for anyone buying leads or calls, and it is verifiable from four separate Google sources.

There is no call asset. Google’s call asset documentation says “Call assets can run on Search campaigns and Smart campaigns.” Demand Gen is not listed. (About call assets)

There is no lead form asset. Google’s lead form documentation says “You can add lead forms to Search and Performance Max campaigns,” and narrows further to ad type: “Responsive search ads creatives are eligible to serve.” (About lead form assets) A second page repeats it: lead forms are “only available for responsive search ads.” (Lead forms in campaigns)

The migration tool strips both on the way in. The Display to Demand Gen error table lists “Click-to-call assets not supported,” telling you to “Add an additional call-to-action (CTA) to the campaign so it can serve on Demand Gen.” It also lists “Lead form assets not supported,” telling you to “Please remove the lead form asset before migrating your campaign.” (Display campaigns move to Demand Gen)

The API agrees. For Demand Gen, Google’s Ads API lists exactly four asset types: TextAsset, ImageAsset, VideoAsset and DemandGenCarouselCardAsset. (Asset creation and usage) No CallAsset. No LeadFormAsset.

The feature comparison against Video Action Campaigns is the cleanest proof. Under extensions, Demand Gen gets product feeds, the call-to-action extension and sitelinks. Video Action Campaigns get those three plus lead form ads (beta) and affiliate location assets (beta). The page’s own legend states that red text marks “a feature not available in Demand Gen.” (Video Action Campaigns upgrade)

The Demand Gen FAQ closes it: “Sitelink extensions are available for Demand Gen campaigns.”

Worth stating precisely, because the internet is confused about this. There is no Google documentation announcing that lead forms were removed from Demand Gen, and none announcing they are coming. Lead form ads appear to have been a Video Action Campaign beta. Read the present tense only, and check the docs again before you plan around it.

That reframe has consequences you can price. Your page has to carry the offer, the qualification and the capture, on a visitor who was not searching for you thirty seconds ago. Page speed, mobile layout and a click-to-call button that actually dials become campaign infrastructure, not conversion-rate housekeeping.

If your funnel ends in a phone call, Google will not count it

Call reporting does not reach this campaign type. Google says it is “Only available on the Search Network,” where forwarding numbers give you conversion information. (About call assets) Demand Gen serves on YouTube, Discover, Gmail, GDN and Maps. None of that is the Search Network.

So for a pay-per-call operation, Demand Gen produces exactly one native signal: a click to your site. Everything after that is invisible to Google unless you make it visible.

That means the tracking stack is not optional here, it is the bidding system’s food supply:

1. Dynamic number insertion on every destination. It is the only record that a call from this campaign ever existed. Without it you have anonymous traffic and a phone that rings.
2. Offline conversion import back into Google Ads. Smart Bidding optimizes toward what it can see. Feed it nothing and it chases the cheapest click on the cheapest surface. That is exactly the failure mode operators blame on the channel.
3. A verification gate before you count a lead as a conversion. Import unverified form fills and you teach the bidder to buy more of whatever produced them. We wrote up how that gate works in ad fraud and OTP verification in lead generation.

Do this before launch, not after. Wire the offline import, confirm a test conversion lands in the Google Ads interface, then launch. An account that spends three weeks before the import works has spent three weeks training the bidder on the wrong signal. Demand Gen also ports 42 days of history when it migrates, so bad learning travels.

Health, insurance and finance get a different campaign than the docs describe

This is the sharpest section in the guide, and the one no competing guide covers.

Google applies two rules to sensitive categories in Demand Gen, on the same page. Read them together.

The first is an inventory rule: “Ads in sensitive categories are generally restricted to YouTube inventory and are not eligible to serve on Gmail or the Google Display Network. However, most sensitive categories, excluding gambling, alcohol, and political content, may serve on the Discover feed if they do not target personalized audiences or predefined Google audiences.”

The second is a targeting rule, four lines later: “Sensitive category campaigns must exclusively target predefined Google audiences, such as Affinity or In-Market audience segments.” (About Demand Gen campaigns)

The Discover exception requires that you do not target predefined Google audiences. The targeting rule requires that predefined Google audiences are the only thing you target. Both cannot hold at once. Our reading is that in practice what is left is YouTube, and we flag it as a reading because Google has not reconciled the two publicly.

Google’s advertising policy tightens it further. “Discovery and Demand Gen campaigns use advertiser-curated audiences by default and may be restricted from serving if targeting products and services that fall within sensitive interest categories.” Separately, advertisers in those categories “are unable to use advertiser-curated audiences.” (Restricted targeting in personalized advertising)

Advertiser-curated covers Customer Match, your own data segments, audience expansion and Lookalike segments.

Read that list again. Lookalike segments are the feature Google itself markets as Demand Gen’s differentiator. Google describes the campaign type as letting you “target unique audience segments such as Lookalike segments,” while Performance Max offers “no hard targeting.” In a sensitive vertical, the reason to choose Demand Gen is the thing you are not allowed to use.

What counts as sensitive. Google’s list runs to 21 interest categories and includes Health and Negative financial status. Separately, the US and Canada carry three “access to opportunities” categories: Consumer finance, Employment and Housing.

What does not count, and this is where operators over-correct. Consumer finance is exemplified as “Credit cards and loans, Banking and checking accounts, Debt management products.” (Consumer finance in personalized advertising) Insurance is not among the examples. Do not assume auto or life insurance is automatically restricted.

Where the consumer finance rules do apply in the US and Canada, they block targeting on gender, age, parental status, marital status and ZIP code. Radius targeting must cover “at least 1 km.”

Health insurance carries a certification gate on top of all of it. Google states: “In the United States, you must be certified by Google in order to advertise health and medical insurance coverage, with the exception of government advertisers, who will be pre-approved.”

Advertisers selling ACA-compliant plans “are required to obtain an additional certificate in order to promote and bid on ACA health insurance related keywords.” US health insurance advertisers are also “required to be certified with G2RS before submitting an application.” (Health insurance policy)

The named scope includes short-term insurance, limited-duration insurance and fixed indemnity health insurance. That is the under-65 off-exchange product set exactly. Only dental, vision and travel health coverage are carved out.

One honest limit: Google says such campaigns “may be restricted from serving.” There is no published rule that says a certified advertiser’s Demand Gen campaign is automatically classified sensitive. Treat classification as a risk to test on a small budget, not a certainty to plan around.

Bidding, and the August 17 change that resets your targets

Demand Gen supports a full set of strategies: “Maximize Clicks, Maximize Conversions, tROAS, Target CPA, Target CPC, and value based bidding.” Google’s ramp advice is to start on Maximize Conversions and shift “to tCPA/tROAS bidding after your campaign received at least 50 conversions.” (Demand Gen FAQ)

Fifty conversions is a real gate, not a formality. If your only importable conversion is a verified call and you get four a week, you are three months from a stable target CPA. That is a planning fact, and it is the first thing to check before you promise anyone a cost per lead.

On change hygiene, Google’s performance guide is specific: “Limit bid adjustments to a maximum of +/- 15% across the campaign duration,” and “slowly flex bids by 5-10% (no more than 15%) per week on campaigns.” (Demand Gen performance guide)

Then there is the dated change. “Starting August 17, 2026, Google will be making changes to its bidding systems.” After that date, “campaigns that are limited by budget that use a target-based bid strategy will more consistently perform toward your bid target, including when you make budget adjustments.”

The worked example is the one that stings. According to Google’s bid strategy change notice, a campaign with a target CPA of $10 that has recently been delivering at half that “will deliver more closely to a $10 actual CPA starting August 17, 2026.”

Demand Gen is explicitly in scope, with an extra warning: “For multi-channel campaigns like Performance Max and Demand Gen, you may also see shifts in how traffic is distributed across different channels.” A Bid Target Adjustment Tool became available July 6, 2026. Google says it “will not automatically adjust your bidding targets or budgets.”

The common mistake: reading this as an efficiency loss and doing nothing. If you have been running a budget-capped campaign that beats its target, your target was never real. Your budget cap was doing the work.

After August 17 the target does the work instead. So set the target to the cost you actually accept per verified unit, before the date rather than after a month of drift.

Creative: what the specs demand and what the surfaces reward

Demand Gen creative is the campaign. There is no keyword to hide behind and no headline in a search result doing the qualification for you.

Source: Demand Gen asset specifications

Two numbers in that table decide whether your best asset ever serves.

The first is the ten-second floor. Google states that “Videos less than 10 seconds are ineligible to serve on YouTube In-stream.” The minimum accepted duration is 5 seconds, so a 6-second cut uploads cleanly and then quietly never appears on the highest-intent video surface in the campaign.

The second is 9:16, which Google marks as recommended for YouTube Shorts. If you are running vertical-first verticals like home services, a landscape-only creative set means Shorts inventory sees a letterboxed ad against native vertical content.

There is also a change tax. Google warns that “Making a change to a creative will reactivate policy approvals, which take around 24-48 hours to complete.” In a regulated vertical that review is the difference between launching against a storm event and launching after it.

The tradeoff most operators get wrong: they treat creative iteration like Search ad copy, changing something weekly. In Demand Gen every edit costs up to two days of serving. It also re-enters the learning the system just finished. Build a proper set at launch, then change deliberately.

If you want to lean on Google’s own generation tools instead, we covered what shipped in the Demand Gen AI creative rollout. Feed-driven creative is in Demand Gen data feeds beyond Merchant Center.

What you can measure, and where the reporting goes dark

Start with billing, because it changes what a cost per lead even means. Google bills Demand Gen differently by surface: “YouTube video: CPM,” “Gmail: Teaser Click,” “Discover feed image: CPC.” (Demand Gen FAQ)

One campaign, one budget, two pricing models running at once. A blended cost per lead out of Demand Gen is an average across CPM and CPC buying. That is why the number moves when the channel mix moves and nothing about your funnel changed. We broke the CPM side down in the Demand Gen CPM billing audit.

Reporting gives you network, not placement. You can “segment your reports by the different channels, using segment by network: Discover, Gmail, YouTube, Google Display Network, Maps.”

But the “Where ads showed” report only breaks out YouTube. Everything else lands in a Total row that “Includes all other channels such as Gmail, Discover, and Google Display Network aggregated together.” (Demand Gen metrics and reporting)

Then there is the measurement gap that matters most to anyone buying calls. Demand Gen can optimize on view-through conversions. That beta is “currently supported for YouTube and Discover Feed only.” It is disabled by default on new and existing campaigns, and recommends a 1-day window for primary VTCs. A view counts “when at least 1 pixel of an ad is on-screen for any time.” The attribution hierarchy is “clicks, engagements, and then views.” (View-through conversion optimized bidding)

The same document states the part that should change how you report to a client: “impression based metrics like Engaged-view conversions (EVC) and VTC are not exported.”

There is a reporting column built for the comparison, Conversions (Platform Comparable). It isolates Demand Gen from the rest of Google, gives the last Demand Gen campaign in the path full credit, and folds VTCs in.

Read Google’s caveat carefully: “This column provides an alternative reporting view and doesn’t impact optimization and bidding.” (Conversions (Platform Comparable)) It is a lens for arguing with a Meta report. It is not what the bidder uses, and it is not the truth about your business.

Three places Google’s own documentation contradicts itself

These are not gotchas. Each one changes a setup decision, and knowing the docs disagree is more useful than picking a side.

On Shorts, do not build a plan on either sentence. The channel picker does list Shorts. Google’s Shorts buying guide does say Demand Gen campaigns “are not a way to target Shorts specifically.” Video view is a campaign type that does document a real Shorts checkbox, and Google frames it for advertisers whose “marketing objective is to influence brand and product consideration and drive video views” (Video view campaigns). That is the wrong instrument for a lead or a call. Treat Shorts as inventory you may receive, never as inventory you can buy on its own.

On exclusions, verify in your own account rather than in the docs. Check whether content exclusions appear for your campaign before you promise a brand-safety posture you cannot deliver, and assume account-level Content Suitability is your real control surface.

On the Discover question, the contradiction sits inside a single page. That is the strongest argument for treating sensitive-category Demand Gen as a YouTube buy and being pleasantly surprised if Discover serves.

Google’s two pages used to disagree on budget. They no longer do. Both now recommend a Demand Gen budget of “at least 10 times your target CPA” for campaigns using target CPA bidding (About Demand Gen). On Maximize conversions that page sends you to its target CPA scaling simulator instead. Under-funding a Demand Gen campaign does not produce a slower version of the same result, it produces a campaign that never exits learning.

Vertical by vertical: where Demand Gen works, and where it does not

The cost column below is deliberately labeled. There is no authoritative Demand Gen cost-per-lead benchmark by vertical. The figures are LocaliQ’s 2026 search advertising benchmarks, drawn from Google Ads search campaigns across thousands of accounts, and they are here as a reference point for the alternative channel, not as Demand Gen costs. (2026 search advertising benchmarks)

U65 and ACA health is the honest no. Short-term, limited-duration and fixed indemnity products are named in Google’s health insurance policy. So certification applies, plus a second certificate for ACA keywords. Health is also a sensitive interest category. That strips Customer Match, your own data segments, audience expansion and Lookalike segments.

Gmail and the Display Network are ineligible. The Discover exception cancels itself out. What remains is YouTube video against affinity and in-market segments. That is an awareness buy, not a lead channel.

The one thing you do control is timing. Marketplace open enrollment starts November 1. December 15 is the deadline for coverage starting January 1. Enrollment closes January 15, after which a qualifying life event is required. (Marketplace dates and deadlines)

Run search and call campaigns for the demand. Use Demand Gen only as an above-funnel YouTube layer feeding a verified page.

Auto insurance is the best fit on this list, and here is why. It is not a sensitive interest category. Insurance is also not among Google’s consumer finance examples. So the audience tooling that makes Demand Gen worth buying stays available: Customer Match, Lookalike segments, custom segments and life-event targeting.

The buying trigger is genuinely interruptible too. Nobody searches for a new carrier on a normal Tuesday. A renewal notice, a rate increase or a new vehicle turns a browsing homeowner into a switcher. That is precisely what an interest-based visual channel is good at.

The constraint is the same as everywhere else. There is no call asset, so this is a form-first delivery to your own page. Run search or call campaigns for the phone product. Our auto insurance leads page covers how carrier-fit filtering works downstream.

HVAC works fully and fails in one specific way. It is not a sensitive category, so every surface and every audience type is available. Maps is the interesting one. A promoted pin is a genuine call source, and it only serves when a non-affiliated location extension is enabled on the account.

The failure mode is pooling. A Maps call and a Display form fill counted under one target CPA are not the same unit. Marginal cost optimization will find the cheap one and starve the good one.

Split the surface into its own ad group with its own target. See the HVAC lead and call verification page for how the downstream unit is defined.

Solar’s problem is time, not policy. It is not sensitive. Financing-led creative can still pull you toward consumer finance restrictions. In the US and Canada those block age, gender, marital status, parental status and ZIP targeting, and require radius targeting of at least 1 km.

The real issue is that a solar decision takes longer than the window you are bidding on. With a recommended 1-day view-through window, Demand Gen will look worse than it is early. On a long window it will look better than it is.

The creative that works is bill math, not environmental framing. The number is available. US residential electricity averaged 18.44 cents per kilowatthour in May 2026, up from 17.37 cents a year earlier. (EIA Electric Power Monthly) Instrument offline conversion import before you spend, or you will judge this channel on the wrong window. More on the downstream unit at solar lead generation.

Roofing is a burst channel pretending to be an always-on one. Non-sensitive, all surfaces available, and storm timing is the entire game. The creative that earns the click is a damage-assessment or claim walkthrough, not a company introduction.

What kills roofing accounts here is leaving the campaign running between events. There is no frequency cap and no campaign-level placement exclusion. So an always-on roofing campaign burns budget against stale inventory when nobody has a wet ceiling.

Run it as a burst, and accept that lead quality is bimodal by design. Context on the unit at roofing lead generation.

Life insurance works, and the risk is self-inflicted. Life insurance is not itself a sensitive interest category. Health is. Google defines Health broadly enough that creative touching medical qualification, conditions or end-of-life framing can pull the campaign into it. You would then lose Customer Match and Lookalikes.

Write to the sensitive-category line even though you are not obliged to. Keep the framing on protecting dependents. Life-event segments like a new home or a new child are the strongest targeting available. Downstream detail at life insurance lead generation.

What to run instead when Demand Gen is not the answer

Demand Gen is an interest channel. It is very good at reaching people who were not looking for you, and structurally bad at capturing them.

If the product is a phone call, run Search with call assets and call reporting. That is the only place Google gives you a native phone conversion with forwarding numbers, and it is not close.

If the demand is emergency home services, Local Services Ads and Search own the moment of need. Demand Gen belongs to the shoulder season, when there is no query to buy.

If you are choosing between automated campaign types, Demand Gen and Performance Max solve different problems. Google’s own comparison puts Performance Max at “no hard targeting” with audience signals as a steer, and Demand Gen at real audience targeting including Lookalike segments. We laid out the full selection logic in Performance Max alternatives for lead generation.

Decision rule: if your acceptable cost is defined per verified call, start on Search. If it is defined per qualified form lead and you have a landing page that can carry the whole funnel, Demand Gen is a legitimate buy. If you cannot import an offline conversion yet, it is neither, and the fix is the tracking stack, not the campaign type.

The pre-launch checklist

Work through this in order. Each item blocks the next one.

1. Classify the campaign before you build it. Is your product in a sensitive interest category, in an access-to-opportunities category, or neither? That answer decides which surfaces and audiences exist for you.
2. Confirm certification. Health and medical insurance in the US requires Google certification and G2RS, and ACA keywords require a second certificate.
3. Decide the channel split by ad group. All Google channels to start. Isolate a surface only with its own ad group and its own target CPA.
4. Turn on location assets if Maps matters. Maps will not serve without a non-affiliated location extension on the account.
5. Put dynamic number insertion on every destination. There is no call asset and no call reporting. This is the only record a call happened.
6. Wire offline conversion import and test it end to end. Confirm a test conversion appears in Google Ads before spend starts.
7. Set a verification gate before import. Import verified units only, or you train the bidder on junk.
8. Build the full creative set at launch. Include 9:16 for Shorts and keep video at or above 10 seconds if you want in-stream. Budget 24 to 48 hours for policy re-review on any later change.
9. Fund the budget at ten times your target CPA. That is what both Google pages now say, and it is scoped to target CPA bidding. On Maximize conversions, Google points you at its target CPA scaling simulator instead.
10. Reset targets ahead of August 17, 2026. If a budget cap has been holding your effective CPA below your stated target, state the target you actually want.
11. Record your baseline before launch. Cost per verified lead and cost per verified call, from your own system, on the day you start.

Two ledgers, and only one of them is yours

Running Demand Gen for leads and calls means living with two incomplete accounts of the same money.

Google’s ledger contains conversions it will not export to you. Engaged-view and view-through conversions steer the bid and never reach your backend, and a view can be one pixel on screen for any length of time. Your ledger contains calls Google never saw, because call reporting does not run on this network, and form fills Google only learns about if you send them back.

Neither ledger is complete, and they will not reconcile. The mistake is trying to make them agree, or picking whichever one is flattering this month.

Run the campaign on the ledger you own. Your verified calls, your verified leads, your cost per unit, measured in your own system, and fed back to Google as the only signal you want it to chase. Google’s numbers are how the auction talks to itself. Yours are the ones you get paid on.

Frequently Asked Questions

What are Google Ads Demand Gen campaigns?

Demand Gen is Google’s campaign type for buying visual, interest-based inventory. Google describes it as serving “on YouTube, Discover, Gmail, and the Google Display Network.” Channel controls sit at the ad group level and also include YouTube Shorts, YouTube in-feed, YouTube in-stream and Maps. It replaced Discovery campaigns, absorbed Video Action Campaigns by April 2026, and is now the only place new Google Display campaigns can be created.

How is Demand Gen different from Performance Max?

The difference is targeting control. Google’s own comparison describes Performance Max as offering “no hard targeting,” where you steer the system with new-customer goals, value rules and audience signals. Demand Gen instead lets you “target unique audience segments such as Lookalike segments,” which are exclusive to Demand Gen. Performance Max also runs across Search and Shopping inventory and supports lead form assets. Demand Gen does not.

Does Google Demand Gen have lead form assets?

No. Google’s documentation states that lead forms can be added to “Search and Performance Max campaigns” and that the feature is “only available for responsive search ads.” The Display to Demand Gen migration tool returns the error “Lead form assets not supported” and instructs advertisers to remove the asset before migrating. Google’s Ads API lists only TextAsset, ImageAsset, VideoAsset and DemandGenCarouselCardAsset for Demand Gen. Every lead must be captured on your own landing page.

Can you run call ads in Demand Gen?

No. Google states that call assets “can run on Search campaigns and Smart campaigns,” and Demand Gen is not listed. Call reporting is described as “only available on the Search Network,” which Demand Gen does not serve on. The migration tool returns “Click-to-call assets not supported.” If you buy phone calls, dynamic number insertion on your landing page is the only record a call from this campaign exists, and you have to import those calls back as offline conversions for bidding to see them.

How much budget does a Demand Gen campaign need?

Both of Google’s pages now give the same answer. The About Demand Gen page says that for campaigns using target CPA bidding it “recommends setting a budget that’s at least 10 times your target CPA,” and the performance guide says “your budget should be at least 10 times your target CPA.” That multiple is scoped to target CPA bidding. On Maximize conversions the About page sends you to its target CPA scaling simulator instead. Google also recommends starting on Maximize Conversions and moving to target CPA only after the campaign has recorded at least 50 conversions, which for a low-volume call product can take months.

Does Demand Gen work for lead generation?

It works when your landing page can carry the entire funnel and you can import offline conversions, because the only in-ad capture Google documents for it is a QR code on connected TV. It works poorly where the product is a phone call, since there is no call asset and no call reporting. It works barely in sensitive categories such as health insurance, where inventory collapses to YouTube and the audience tools that justify choosing Demand Gen are unavailable. Auto insurance, HVAC, roofing and solar are the stronger fits.