Demand Gen vs Performance Max: Which Google Ads Campaign Fits Your Funnel

Half the accounts I look at are paying for demand someone else created, and the other half are creating demand that nothing in the account captures. Google presents Demand Gen and Performance Max as two settings on the same dial. They are not. Demand Gen buys attention on YouTube, Discover, Gmail, and the Display Network, where nobody has typed anything into a search box. Performance Max buys that same attention plus the Search and Shopping inventory where intent already exists. One campaign creates demand. The other captures it. Once you see that split, the question stops being "which campaign is better" and becomes "which job is my account not doing yet".

Demand Gen Google Ads: What the Campaign Type Actually Is

Google describes Demand Gen as its campaign type for social-style, visual, video-first advertising, and that framing is accurate. It reaches up to 3 billion monthly active users across Google's own surfaces: YouTube (in-stream, in-feed, and Shorts), Google Discover, Gmail, Google Maps in beta, and the Display Network, which is currently migrating into Demand Gen rather than running beside it. You serve image, video, and native formats, and you can preview and edit creative before it goes live. Demand Gen ads are assembled per placement, which is why a single campaign needs separate YouTube assets for in-feed, in-stream, and Shorts.

Two controls matter more than the surface list. Since an open beta in June 2025, Demand Gen supports channel-level opt-out, so you can exclude Discover or Shorts without abandoning the campaign. And Demand Gen is the only Google Ads campaign type that supports lookalike audiences, which is why performance teams moving budget over from Meta find the interface familiar.

Audience signals work the same way they do in Performance Max: as hints, not restrictions. You build custom segments, upload your own data, and add interest or lookalike segments to tell the algorithm where to start. Google then expands beyond them. If you were hoping for hard targeting, this will disappoint you daily.

One piece of history is worth knowing. Video action campaigns were retired and replaced by Demand Gen from July 2025. If your account still carries VAC-era creative and naming, that is not a rename you can leave alone. As for scale, Google reports that Demand Gen campaigns including TV screens drive an average of 7% additional conversions at the same ROI through shoppable CTV, and that advertisers who moved Display campaigns into Demand Gen see an average 9.5% ROI increase (both figures are Google's own vendor data, published in its January and May 2026 announcements, not independent measurement).

What Performance Max Actually Is, and Where It Differs Structurally

I covered Performance Max campaigns in detail in the sibling guide, so here is the short version. Performance Max buys every Google surface from a single campaign: Search, Shopping, Display, YouTube, Discover, Gmail, and Maps. You supply asset groups (headlines, descriptions, images, logos, video), a conversion goal, and audience signals, and the algorithm decides placement, audience, and price. Conversion-value bidding is the default posture.

The structural difference between the two campaign types is not the surface list, which overlaps heavily. It is that Performance Max can capture intent and Demand Gen cannot. Performance Max includes Search and Shopping inventory. Demand Gen never touches a search results page. That single fact decides most real accounts: Performance Max is where existing demand gets harvested, and it is also where your proven exact-match queries can quietly get absorbed if you do not exclude them.

Reporting transparency is the other divergence, and it cuts against Performance Max. Asset-group and placement reporting in Performance Max is deliberately thin. Demand Gen is not generous either, but the arrival of view-through conversions in Demand Gen reporting at least gives it the measurement shape that paid-social teams already trust.

The Funnel-Position Split: The Real Demand Gen vs Performance Max Decision

Here is the part most comparisons skip. Demand Gen and Performance Max are not competing for the same impression. They are competing for the same budget line, which is a different problem.

Demand Gen exists to manufacture the intent that Performance Max later harvests. Lunio's strategy team put it about as cleanly as it can be put: Demand Gen is there to create intent, Performance Max is there to capture it. When each does its own job, both get cheaper.

The practical split for an account running both: Performance Max owns conversion capture, including Shopping for ecommerce and the queries your Search structure cannot cover. Demand Gen owns prospecting and remarketing above that line. Decide which campaign owns prospecting versus remarketing in writing, and use exclusions so the two do not chase the same users. Without that rule, Demand Gen will happily retarget the exact remarketing audience your Performance Max campaign is already paying for.

When Demand Gen Beats Performance Max (and When It Does Not)

Demand Gen wins on four conditions. Creative is video-first or visual-native and your team wants real control over it. The job is prospecting, not harvesting. You need lookalike or first-party-signal audiences, which Performance Max does not give you. Placement control matters, whether that means excluding Shorts or running Display Network only.

Then there is the account that had no choice. If Display was already in your mix, the migration settled the question for you. Google's announced timeline has voluntary migration open from June 2026, new Display-only campaigns eventually stopped, and the transition complete by 2027.

Performance Max wins on other conditions, and it wins them decisively. Conversion volume and value data are mature. Shopping or Search inventory is the point, and the user is already searching. Or you accept the reduced transparency in exchange for reach across everything.

Now the honest part, because no comparison page seems to include it. Demand Gen dies quietly on creative fatigue. It is a video-first format, and the same three assets will stop working long before your budget does. It also needs enough money to leave the learning phase at all. Google's own recommendation, repeated by Search Engine Land, is roughly 100 pounds per day per campaign or 20 times your target CPA, whichever is higher. Below that, the signal flow stalls and you conclude the campaign type is bad. It is not bad, it is underfunded. And because Demand Gen has no Search inventory, it cannot rescue a broken conversion path. If your landing page does not convert, Demand Gen will simply deliver more people to a page that does not convert.

Performance Max fails differently. Its reporting hides which asset and which placement took the credit. It can serve brand-search traffic and flatter its own performance with conversions your brand campaign would have won anyway. And in a data-thin account, it has nothing to learn from. I have watched accounts move 100% of Google spend to Performance Max in month one and blame the algorithm in month two, when the actual problem was 14 recorded conversions a month.

One warning that applies to both. In a B2B lead-gen account, Performance Max is not automatically the answer. Agency export data from 42 Agency covering 2022 to 2026 found Search returning 553% ROAS against Performance Max at 436% in B2B lead gen, and exact-match keywords delivering a cost per MQL roughly half that of phrase match ($1,200 versus $2,800). That is agency data, not a Demand Gen metric, but it is a useful corrective to the belief that the newest campaign type always wins.

Demand Gen Specs: Structure, Assets, and the Feed Question

Structure follows the Performance Max model. You build asset groups, each a themed bundle of images, videos, logos, headlines, and descriptions, and each representing one product line or one message. Audience signals attach to the asset group, and so does the campaign's budget. The rule that trips people up is the same one from Performance Max: signals are hints. Google expands past them.

Demand Gen specs are where most guides stop, and where the campaign actually gets built. Verified against Google Ads Help's Demand Gen creative specifications (checked September 2026):

Two entries in that table are the ones you cannot guess: the 10-second floor for YouTube In-stream and the 9:16 Shorts recommendation. Getting either wrong produces assets that upload fine and never serve.

One operational constraint that catches almost everyone: video assets for Demand Gen must be uploaded to YouTube first and referenced by URL. You cannot drop an MP4 into the campaign and launch.

On text assets, be careful with any number you read. The headline, long headline, and description character limits most articles quote (30, 90, 90, business name 25) come from the Performance Max Help page. Demand Gen uses the asset-based system with its own limits, and I have not been able to verify Demand Gen's own text caps from Google's documentation. Treat those numbers as Performance Max limits until you confirm them on the Demand Gen page for your account.

The product feed question, answered correctly

You will read that Demand Gen does not use a Merchant Center feed. That was true, and it is not true anymore. Google added product feed integration for Demand Gen in early 2025, and its Display-migration matrix lists product feeds as partially supported in Demand Gen, with some business data feed types still pending.

The accurate difference is the role of the feed, not its existence. In Performance Max and Shopping, the feed is the campaign: your catalog drives inventory, eligibility, and the ads themselves, and without a healthy feed there is nothing to show. In Demand Gen, the feed is an option. It powers shoppable product cards and carousels next to your lifestyle creative, and the campaign still runs on signals and creative with no feed at all.

So the decision input is this. Performance Max has a feed dependency. Demand Gen has a feed option. If your retail catalog is the point, Performance Max is where it belongs. If your videos are the point and the catalog is a bonus, Demand Gen is the fit.

Managing Demand Gen Campaigns at Scale

This is where Demand Gen gets hard, and it is the reason most accounts plateau after a good first month.

In practice, demand gen campaigns run on creative and signals, not keywords, which means the levers you have for Search campaigns do not exist here. You cannot fix a stalled campaign by adding negatives or raising a bid on a proven term. You fix it by replacing fatigued creative, tightening or widening audience signals, and cutting spend on surfaces that never convert. That work is ongoing, and it is not visible in a single campaign view.

Three things break first. Creative fatigue is the number one killer, and it shows up as rising CPM with flat conversions before anyone notices. Surface-level spend inefficiency hides inside the aggregate: a campaign can look healthy while most of its budget flows to one placement that has never converted. Audience-signal drift happens when you set signals once at launch and never revisit them.

None of this is visible in the standard report. You need spend broken out by surface, creative performance tracked over time, and signals reviewed on a schedule. That is the job Demand Gen campaign management exists to do, and if you are running Google Ads at any scale, Performance Max management belongs in the same workflow. Both sit on the same reporting layer, which matters because the moment you run both campaign types you are comparing them against one budget. Bringing Ads data together with Google Ads reporting in one place is what makes the split decision measurable rather than theoretical.

Google's own numbers support the investment. In its August 2026 Demand Gen announcement, Google reported that improvements shipped through the second half of 2025 are driving an average 30% increase in conversions or conversion value for advertisers using the campaign type. That is Google's vendor data, and it is the kind of claim you should test in your own account rather than take on faith. StoreGrowers' benchmark synthesis makes the honest point that public Demand Gen data is genuinely thin, so most accounts end up assembling their own baseline from placement mix and objective.

The Demand Gen vs Performance Max Decision Table

If you only keep one thing from this article, keep this:

The Bottom Line

Demand Gen and Performance Max are not competitors. They are the two halves of a funnel that most accounts only fund one of. If your conversions all come from people who were already searching, you are paying for demand someone else created. If your spend all goes to attention and nothing captures it, you are renting reach.

Pick the campaign type by the job, not by the launch date. Fund the one you chose well enough to learn, set written ownership rules if you run both, and watch the creative before you watch the bid.