Using UGC Content in Google Performance Max Campaigns

UGC works inside Performance Max because it solves a problem no amount of studio lighting can fix: it looks real. Google's algorithm handles bidding, placement, and channel mix on its own, but it can't manufacture trust out of a polished product shot. That gap is why UGC has become the raw material serious PMax advertisers build their asset groups around, and why brands still feeding it stock photography are leaving performance on the table.

PMax is a single campaign setup that Google's machine learning spreads across Search, Shopping, YouTube, Display, Discover, Gmail, and Maps, deciding on its own where a given ad performs best. But automation isn't a guarantee of quality. Adalysis ran a study across 3,300 campaigns and found standard Search campaigns typically beat PMax when both bid on the same terms. Over a million advertisers now run PMax worldwide, and a good chunk of them are feeding it the same tired product photography and getting average results back. The algorithm only works with what it's handed. If it is handed something generic, it optimizes generic.

What Makes UGC Better Creative Fuel for PMax

UGC, in the PMax context, means phone-shot reviews, testimonials, demos, unboxings. Cheap to produce, easy to reshape into different aspect ratios, and built to look native to whatever feed it lands in. That last part affects how well the asset performs: phone-shot reviews, testimonials, demos, and unboxings read as native to the feed, which is what lets them succeed where polished content doesn't.

Why does production value take a back seat here? Because of a trust gap that no lighting rig closes. Nielsen puts trust in peer recommendations at 92%, against branded messaging sitting far below it, a gap large enough to shape strategy on its own. That's not a soft preference. The trust gap raises click-through rate, cost-per-click, and watch time in the direction PMax optimizes toward, and it is visible directly in those numbers.

Head to head against traditional brand ads, UGC's advantage isn't subtle. Click-through rates run close to 4x higher. Cost-per-click drops by around half. UGC video holds attention roughly a third longer than polished brand video, and shoppers who encounter UGC convert at a rate more than double what polished content pulls in. None of that is a stylistic preference. It's what happens when an ad looks like something a friend sent you instead of something a brand paid for.

How PMax ingests, assembles, and scores creative assets

Asset groups are where the actual work happens. Think of an asset group as PMax's version of an ad group, a container holding images, video, headlines, descriptions, and logos, which Google's system mixes and matches on its own, without an advertiser choosing the pairings by hand.

The AI tests combinations dynamically. A YouTube placement might pair Video A with Headline 3. A Search ad pulled from that same asset group might run Headline 7 next to Description 2 instead. Nobody signs off on that combination before it runs.

Ad Strength is the diagnostic tool built to flag when an asset group lacks enough variety, across text, image, and video, to give the algorithm real options to test. Treat a weak score as a warning.

One detail trips up more advertisers than it should. If no video gets uploaded to an asset group, the system may auto-generate one out of whatever static images are sitting there. That auto-generated version is rarely what an advertiser would choose to run on YouTube and Shorts. Uploading a real video, shot with intent, is the only way to keep that decision in-house.

Technical specifications UGC must meet to run across every PMax channel

PMax draws from one shared pool of assets to build ads across Search, Display, YouTube, Gmail, Discover, and Maps. Missing a spec causes an asset to quietly lose eligibility for entire placements, with no obvious warning that it happened.

Images: up to 20 total, covering landscape (1.91:1), square (1:1), and portrait (4:5 or 9:16), plus up to 5 logos. Brief every creator to shoot in all three orientations from the start. A vertical-only shoot, which is what most creators default to because that's what performs on short-form social, locks the asset out of Display and Gmail before it ever gets a chance to run there.

Video needs a format compatible with a major video-sharing platform, minimum 10 seconds long. Upload custom versions in all three ratios, horizontal (16:9), square (1:1), vertical (9:16), rather than letting Google auto-crop one version into the others. The hook needs to land in the first five seconds. That's the window before a YouTube viewer can hit skip, and it doesn't forgive a slow build-up.

Text: 15 headlines, 5 descriptions. Write copy that is comfortable next to UGC's casual tone. Corporate phrasing paired with a phone-shot testimonial reads like a mismatch, and it undercuts the exact authenticity the video was supposed to sell.

Sourcing UGC at the volume and quality PMax creative rotation demands

Creative fatigues inside PMax faster than most advertisers expect. Adding a handful of new assets every six to eight weeks is usually enough to keep performance from sliding. A full creative reset isn't required, just a steady swap of whatever's dragging the group down.

That cadence turns UGC sourcing into infrastructure that runs continuously. A brand running PMax year-round needs a pipeline producing new, usable assets every couple of months, indefinitely, with no real end point.

The supply side shifted fast. The pool of UGC creators grew 93% between 2024 and 2025, and 66% of them now offer UGC as a formal service instead of something they do occasionally. More options on paper, but also a lot more noise to sort through before finding someone worth working with repeatedly.

A few paths for sourcing, each with its own tradeoff. Vetted marketplaces like Insense, Billo, or Trend offer pre-screened creators and fixed pricing: Billo focuses on fast turnaround for video ads, while Trend leans toward viral, trend-forward content at a higher price point. Good for fast volume, weaker for hand-picking exactly who shoots for the brand. Aspire is at the enterprise end, with creator discovery aided by one model starting around $1,000, built for brands that need scale and a premium creator network behind it. Direct sourcing, reaching out to people already using the product and tagging the brand organically, produces the most authentic content by a wide margin, though it's slow and hard to scale without real operational effort behind it. Structured campaigns, hashtag challenges or contests tied to a specific brief, generate a batch of content around one theme while building organic reach at the same time.

Handling all of this in-house (sourcing, vetting, contracts, usage rights) adds up to a real operational load, and it's the reason some brands hand it off to an agency that specializes in creator marketing end to end. Whichever route gets picked, the pipeline has to survive past the first content drop, or the six-to-eight-week cadence collapses the moment it gets tested.

Vetting creators and selecting assets for paid media performance, not social appeal

Vetting a UGC creator for paid media isn't the same exercise as vetting an influencer, and treating it like one is a common mistake. Follower count barely matters here. What predicts performance is authentic delivery, a grasp of platform-native trends, and a voice that fits the brand without sounding scripted or read off a card.

A handful of things actually forecast whether an asset performs once it's inside PMax. The first five seconds matter most: does the opening moment stop a scroll on its own, without leaning on a branded intro to do the work? How the creator handles the product matters almost as much. Interacting with it like they'd actually use it reads differently than reciting lines at it. Format range counts too. Has this creator delivered footage in multiple aspect ratios before, or do they only shoot vertical because that's the only thing they know? And track record with paid campaigns specifically tells you more than organic likes and comments, since those measure a different kind of engagement.

Watch for inflated portfolios and padded engagement numbers. A creator pool that grew 93% in a year attracts opportunists right alongside genuine talent, and view counts alone won't tell those two apart. Ask for references or real campaign data instead.

Scaling a creator roster faces a ceiling problem: managing that refresh cycle by hand becomes increasingly difficult as the roster grows, and industrializing intake, briefing, and vetting becomes necessary instead. Scaling a creator roster to hit that cadence means building repeatable systems around those processes. Brands trying to manage all of that one creator relationship at a time tend to hit that ceiling faster than they expect.

Structuring asset groups so UGC signals reach the right placements

Dumping every piece of UGC into a single asset group is the most common structural mistake, and it quietly caps performance before anyone notices why results plateaued. The algorithm can't tell a shoe testimonial from a jacket demo if they're sitting in the same bucket. It ends up optimizing for the average of everything instead of the best of anything.

A better structure segments UGC along a few lines. By product or SKU cluster: shoe content in one asset group, jacket content in another, with no image or headline combinations crossing over between product lines that don't share an audience. By creative angle: testimonial-style UGC separated from demonstration-style UGC, so the algorithm can learn which format wins on which channel instead of blending the signal into mush. By funnel stage, where the data supports it: broader, story-driven UGC for awareness, sharper claim-specific UGC (before-and-after, a direct call to action) for the conversion push.

Mixing UGC-style and polished production inside the same asset group is fine, even useful. PMax's machine learning routes each toward where it performs best. UGC tends to win on mobile feeds and Shorts. Polished video tends to hold up better on connected TV and desktop in-stream. Letting the algorithm sort that out beats guessing manually, every time.

Pairing UGC asset groups with first-party audience signals, Customer Match lists, custom segments built around search intent, sharpens the algorithm's aim further. Someone who's curious about a product and someone who's ready to buy it are not the same audience, and treating them as one flattens a distinction PMax is fully capable of using.

Testing UGC assets inside PMax and iterating based on what the algorithm surfaces

PMax's built-in A/B testing feature expanded to all PMax campaigns in January 2026, and it's the right tool for structured comparisons. Testing UGC testimonial against product demo, or UGC against polished video, directly through that feature beats trying to guess from blended, campaign-wide metrics that average everything together.

A newer reporting addition, the "Ads Using Video" segment, lets advertisers isolate performance from ads that included video versus those that didn't. That's a distinction PMax reporting hasn't offered before without restructuring an entire campaign just to force the comparison manually.

Check asset-level performance scores on a regular schedule. When one asset is clearly dragging, swap that one out rather than refreshing the entire group. A smaller change disturbs less of what the algorithm has already learned about what works, while still keeping the rotation from going stale.

Judging new assets too early skews the results, because the numbers being read back during the learning phase haven't settled yet. Give new assets roughly two to three weeks after upload before drawing conclusions from their scores. The learning phase needs that runway to settle. Judge too early, and the numbers being read back are still noise, not signal, no matter how convincing they look on day four.