Google Shopping’s AI-Powered Search Ads Are Reshaping DTC Acquisition Math

For DTC operators who built their customer acquisition playbooks around tightly controlled Google Shopping campaigns, the summer of 2026 has been a reckoning. Google’s continued expansion of Performance Max — now accounting for an estimated 68% of all Shopping ad spend on the platform, according to data from Tinuiti’s Q2 2026 Benchmark Report — has fundamentally altered how brands buy traffic, attribute conversions, and forecast CAC.

The core tension: Performance Max hands Google’s AI nearly full control over where, when, and how ads appear. For some categories — home goods, pet supplies, beauty accessories — the results have been exceptional. For others, particularly fashion and high-consideration electronics, merchants are reporting ROAS volatility they can’t explain with existing attribution tools.

What Is Driving the Performance Max Volatility in 2026?

The short answer, according to agency leaders who manage nine-figure Google ad budgets, is signal dilution. Performance Max ingests first-party data, product feeds, creative assets, and audience lists, then optimizes across Search, Shopping, Display, YouTube, Discover, Gmail, and Maps simultaneously. When those signals conflict — or when a brand’s first-party data is thin — the algorithm tends to over-index on bottom-funnel placements that cannibalize organic traffic rather than sourcing net-new customers.

“We had a home décor client spending $180K a month in PMax and their new customer acquisition rate had dropped to 31% of total conversions. The algorithm was essentially retargeting their email list at Shopping CPCs. That’s not acquisition — that’s an expensive loyalty program.” — Cody Plofker, CMO at Jones Road Beauty and growth advisor to several Shopify mid-market brands

Plofker’s observation is increasingly common. Brands that haven’t segmented their PMax campaigns — separating prospecting asset groups from retargeting signals — are effectively subsidizing Google’s placement engine with high-intent bottom-funnel users while reporting inflated ROAS numbers that don’t hold up against MER (marketing efficiency ratio) analysis.

Which Merchant Categories Are Winning and Losing on Google Shopping Right Now?

The category divide is stark. Tinuiti’s Q2 data shows that home and garden verticals averaged a 6.2x ROAS on Performance Max placements, while fashion apparel averaged 2.9x — a gap that’s widened by roughly 40 basis points since Q4 2025. Pet supplies and consumable beauty both came in above 5x.

Agency operators attribute the fashion underperformance to two structural issues:

• High SKU count and frequent inventory turnover make feed quality management extremely difficult at scale, leading to suppressed impressions on high-margin items
• Google’s visual matching algorithm still struggles with apparel fit, color accuracy, and style differentiation — driving click-through rates that look healthy but convert poorly
• Return rates in fashion (averaging 28-34% across DTC brands surveyed by Narvar in 2026) erode actual ROAS well below what Google’s last-click or data-driven attribution reports

High-consideration electronics face a different problem: comparison shopping behavior means users click multiple Shopping ads before converting, and Google’s attribution model — even in data-driven mode — tends to overcredit the last Shopping placement, masking the true cost of the research journey.

How Are Sophisticated Operators Restructuring Their Shopping Campaigns?

The operators pulling the best numbers in 2026 are running what Nik Sharma, founder of Sharma Brands, calls a “segmented feed architecture” — a structure that separates products into distinct Performance Max campaigns by margin tier, inventory velocity, and new-customer probability.

“The brands killing it right now are treating their Google feed like a merchandising floor. They’re not dumping 4,000 SKUs into one PMax campaign and letting Google decide. They’re running three to five campaigns with clear signals: here’s my hero product, here’s my new customer target ROAS, here’s the customer list I explicitly want to exclude.” — Nik Sharma, Sharma Brands

Practically, this looks like:

• A “hero SKU” campaign capped at 50-80 products with a 500-700% target ROAS and new customer acquisition goal enabled
• A “margin defense” campaign for high-AOV products with customer match lists of existing buyers uploaded as negative audience signals (using the “bid lower” setting since explicit exclusions remain unavailable in PMax)
• A clearance campaign with no ROAS floor, optimized purely for revenue velocity to clear aged inventory
• Separate asset groups for prospecting creative (UGC video, lifestyle imagery) versus retargeting creative (product-focused, price-prominent)

Feed quality remains the upstream bottleneck. Operators at brands like Caraway Home and Mejuri — both of whom declined on-record comment — have reportedly invested in dedicated feed management tooling. Platforms like DataFeedWatch, Feedonomics, and GoDataFeed have seen meaningful adoption growth in 2026 as merchants recognize that title optimization, custom label hierarchies, and real-time inventory sync are now table stakes, not differentiators.

What Role Is AI Playing in Google’s Search Ad Targeting Beyond PMax?

Beyond Performance Max, Google’s broader AI push is reshaping how Shopping ads surface in traditional Search results. The rollout of AI Overviews — now appearing on an estimated 46% of commercial queries in the U.S. according to SparkToro’s June 2026 tracking data — has compressed organic click-through rates for product category pages by as much as 22% year-over-year, according to analysis from Conductor.

That organic compression is pushing incremental spend toward Shopping placements, which now appear both above and alongside AI Overviews on high-intent queries. The net effect: CPCs for top Shopping placements have risen 14-19% across most categories since January 2026, per Skai’s Q2 2026 Ecommerce Index.

“Everyone’s CAC is going up and they’re blaming Meta. But a meaningful chunk of that increase is Google Shopping CPCs. AI Overviews ate their organic traffic and now they have to buy it back at auction.” — Melissa Berends, VP of Paid Search at Logical Position, speaking at eTail West 2026

The practical implication for DTC operators: brands that invested in programmatic SEO and long-tail content as a hedge against paid spend are seeing that hedge partially erode. The counter-strategy gaining traction is aggressive Shopping ad coverage on long-tail, high-intent queries — terms like “best non-toxic ceramic pan set under $200” — where AI Overviews are less likely to dominate and Shopping carousel placements convert at above-average rates.

How Should DTC Brands Think About Attribution Across Google’s AI-Driven Placements?

Attribution remains the most contested operational question. Google’s native data-driven attribution model is increasingly viewed with skepticism by operators who cross-reference it against triple-whale, Northbeam, or Rockerbox multi-touch data. The delta between Google-reported ROAS and blended MER is widening for brands running heavy PMax budgets — sometimes by 30-40%.

The current best-practice stack, according to a half-dozen agency leaders interviewed for this article:

• Run Northbeam or Triple Whale as the source of truth for channel-level efficiency, calibrated against 7-day new customer revenue
• Use Google’s “new customer acquisition goal” with a 10-15% new customer value uplift to bias the algorithm toward prospecting
• Cross-reference Google Analytics 4’s session-based data against platform-reported conversions weekly — gaps above 20% signal significant view-through or cross-device attribution inflation
• Run monthly incrementality tests using geo holdouts (Measured.com has built a self-serve version of this now) to validate whether Shopping spend is driving genuine lift

The incrementality question is increasingly urgent as Q4 budgets get locked in. Several agency operators told Ecommerce Times they are advising clients to hold back 8-12% of their Google Shopping budgets for October incrementality testing before committing Q4 peaks — a conservative posture given that Q4 2025 saw CPC spikes of 28-35% in peak weeks.

What Should Operators Do Before Q4 2026 to Protect Their Google Shopping ROAS?

With peak season roughly ten weeks out, the operational checklist from agency leads and in-house growth teams is converging around a few non-negotiable actions:

• Audit your product feed title structure. Google’s algorithm increasingly uses title relevance as a proxy for bid eligibility. Front-load brand, product type, and key attributes in the first 70 characters.
• Segment PMax campaigns by margin and customer intent. Do not run new-customer prospecting and retargeting within the same asset group if you want interpretable data.
• Upload your suppression lists now. Customer match lists take up to 48 hours to fully populate. Get existing buyers into the system so you can apply the bid-down signal before Q4 CPCs spike.
• Set up a Merchant Center feed rule for price competitiveness. Google’s Price Competitiveness report in Merchant Center identifies SKUs where your price is suppressing impressions — fix those before October.
• Test Shopping campaigns against Standard text Search campaigns on your top 10 hero queries. Some operators are finding Standard Search outperforms PMax Shopping on branded and near-branded terms at 30-40% lower CPC.

The broader strategic reality is that Google Shopping has become a more expensive, more opaque, and more essential channel simultaneously — a combination that demands operational precision rather than set-and-forget campaign management. For DTC brands heading into what most forecasters expect to be a competitive Q4 2026, the operators who’ve done the feed work, the segmentation work, and the attribution calibration work will have a measurable structural advantage over those still running single-campaign PMax setups.

The AI is not going away. The question is whether your feed, your signals, and your measurement infrastructure are good enough to make it work for you rather than against you.