Google Shopping’s New AI-Powered Bidding Is Splitting DTC Ad Budgets

Google’s mid-May rollout of enhanced AI bidding signals inside Performance Max — specifically the new “Demand Forecasting” layer that adjusts Shopping bids in near-real-time based on search trend velocity — is landing inside DTC ad accounts at a disruptive moment. Brands that built their acquisition stack around stable Google Shopping CPCs are now seeing cost-per-click volatility they haven’t experienced since Smart Shopping campaigns were first forced onto the platform in 2022.

For Shopify merchants running between $500K and $10M in annual revenue, the shift is measurable. According to data from Northbeam, which tracks blended ROAS across roughly 1,800 DTC brands, Google Shopping ROAS dropped an average of 14% in the two weeks following the May 12 rollout, while brands that quickly restructured their asset groups and audience signals recovered to within 6% of prior benchmarks by week three.

What exactly changed in Google’s May 2026 Performance Max update?

The update — Google refers to it internally as “PMax Signal Expansion 3.0” — does two things that are operationally significant for ecommerce advertisers. First, it expands the demand forecasting model to incorporate real-time Google Trends data at the product category level, meaning bids on individual Shopping listings can shift by as much as 30% intraday based on search momentum signals Google’s algorithm detects. Second, it reduces the manual override window for Search theme inputs from 14 days to 7 days, compressing the feedback loop merchants relied on to course-correct campaigns gone sideways.

“The 7-day override window sounds like a small change, but it’s operationally brutal for brands selling anything seasonal or trend-driven. You barely have enough data to make a decision before the window closes and the algorithm locks you out again.”

That’s Cody Plofker, CMO at Jones Road Beauty, speaking at a virtual roundtable hosted by the DTC Growth Alliance on May 21. Plofker, who manages a mid-seven-figure annual Google spend, said his team rebuilt three of six PMax campaigns from scratch after the update pushed Shopping ROAS from 4.1x down to 3.2x in the first 10 days.

Which ad account structures are surviving the volatility best?

Agency operators are converging on a few structural fixes that appear to stabilize performance in the new environment. Tinuiti’s commerce team, which manages Google Shopping for over 200 brands, issued an internal playbook on May 19 that Ecommerce Times reviewed. The core recommendations:

• Segment PMax campaigns by margin tier, not just product category. Brands blending high-margin SKUs with loss-leader products inside a single asset group are seeing the algorithm over-index spend on volume items at the expense of profitable ones.
• Upload at least 15 lifestyle image assets per asset group. The demand forecasting layer appears to weight creative diversity more heavily than previous PMax iterations, and accounts with sparse creative libraries are getting penalized in auction density.
• Run a parallel Standard Shopping campaign for top-20 hero SKUs. Standard Shopping gives manual bidding control that PMax doesn’t, and several Tinuiti clients have shifted 20–30% of their Shopping budget back to Standard as a stabilizing hedge.
• Refresh audience signals weekly, not monthly. The shortened override window means stale audience lists become liabilities faster than before.
• Set target ROAS floors at the campaign level, not the asset group level. Granular tROAS settings at the asset group level appear to confuse the new bidding model during high-volatility periods.

Brett Curry, CEO of OMG Commerce and one of the most followed voices in Google Shopping strategy, posted a detailed breakdown on LinkedIn on May 22 that has since been shared over 3,400 times. His take was blunter than most agencies were willing to put in writing.

“Google has essentially told advertisers that the algorithm knows better than you do, full stop. That might eventually be true. But right now, in Q2 2026, the model is still making enough mistakes on niche product categories that handing over full control is a legitimate business risk.”

Are DTC brands shifting Google Shopping budgets to Meta as a result?

Yes, and the numbers are meaningful. Rockerbox, which tracks cross-channel attribution for roughly 500 DTC brands, shared data with Ecommerce Times showing that from May 13 through May 25, brands in its network increased Meta Advantage+ Shopping Campaign (ASC) spend by an average of 11.4% while simultaneously pulling back Google PMax budgets by 8.7%. That’s not a wholesale channel abandonment — but it represents a real reallocation that’s likely to persist through Q3 if Google Shopping volatility doesn’t stabilize.

The irony is that Meta’s own ASC product has had its own turbulence this year. Several DTC operators who spoke to Ecommerce Times on background noted that ASC CPMs rose sharply in late April — up roughly 18% week-over-week in the beauty and apparel verticals — before settling back down in mid-May. The convergence of Google Shopping instability and Meta CPM spikes created a two-week window in early May where brands running $50K-plus in monthly ad spend had nowhere efficient to put incremental dollars.

“We basically paused top-of-funnel spend for eight days in early May and put the budget into email flows and SMS win-backs instead. Klaviyo and Attentive don’t have algorithm updates that crater your ROAS overnight.”

That quote comes from Ari Murray, VP of Growth at Sharma Brands, who has been publicly vocal about over-reliance on paid channels since early 2025.

How are Google Shopping agencies adjusting their retainer structures?

The volatility is creating a secondary business problem for performance marketing agencies: retainer models built around percentage-of-spend are getting squeezed when brands pull back budgets mid-quarter. Two agency operators told Ecommerce Times that clients invoked force-majeure-style contract language in May to temporarily reduce spend floors, an unusual move that reflects just how abruptly the update disrupted planning cycles.

Some agencies are responding by moving toward hybrid retainer-plus-performance models that include a fixed monthly fee for strategy and account management, with performance bonuses tied to ROAS above a defined floor. For a $200K/month Google Shopping account, the structure might look like a $12,000 flat fee plus a 0.5% performance kicker on every dollar of ROAS above 4.0x — giving both parties skin in the outcome regardless of platform-side volatility.

What does this mean for Google Shopping’s role in the 2026 DTC acquisition stack?

The longer-term question isn’t whether brands will stay on Google Shopping — they will, because intent-based search traffic remains the highest-quality acquisition signal in ecommerce — but whether Google Shopping will continue to serve as the primary acquisition lever or get repositioned as a mid-funnel capture channel while brands diversify spend into TikTok Shop affiliate, connected TV, and owned channels.

TikTok Shop’s affiliate commission model, in particular, is increasingly attractive to brands that are exhausted by paid media volatility. Several Shopify merchants in the $2M–$15M revenue range told Ecommerce Times they’ve shifted 10–15% of what was previously Google Shopping budget into TikTok Shop creator commissions — a model where the brand pays only on conversion, with no upfront CPM or CPC exposure.

Katya Allison, director of marketing at Kiddo Co., a children’s apparel brand doing roughly $8M annually on Shopify, put it plainly:

“When Google changes the algorithm and my CPC goes up 25% in a week, I have no recourse. When a TikTok creator doesn’t convert, I don’t pay. That asymmetry is changing how I think about where to put the next incremental dollar.”

Google has not issued a public comment on the PMax Signal Expansion 3.0 rollout’s performance impact, though a Google Ads liaison acknowledged on the Google Ads Help Community forum on May 20 that “some accounts may experience a temporary adjustment period as the model recalibrates to expanded signals.”

What should Shopify and DTC merchants do right now?

Operators who spoke to Ecommerce Times consistently flagged the same near-term action items:

• Pull a 30-day pre/post ROAS comparison immediately — segmented by campaign type, not blended account performance. Blended numbers hide where the damage is concentrated.
• Audit asset group creative depth. Any group with fewer than 10 image assets and 3 video assets is likely underperforming against Google’s new scoring model.
• Check your tROAS targets against current conversion rates. If your conversion rate dropped 5% post-update but your tROAS target didn’t change, the algorithm is likely withholding spend rather than bidding aggressively.
• Consider a Standard Shopping holdout campaign for hero SKUs generating 40%+ of your Google Shopping revenue. The manual control is worth the management overhead during volatile periods.
• Run an SMS abandoned cart sequence test this week. Klaviyo and Postscript both report higher-than-average abandoned cart recovery rates in Q2 2026 — brands that invested in owned channel infrastructure are less exposed to paid media shocks.

The broader takeaway for DTC operators is structural: platform dependency has always been a risk, but the acceleration of AI-driven auction changes on both Google and Meta means that variance in paid media performance is now a permanent operating condition, not a periodic disruption. Brands that build margin buffers, own their customer data, and maintain diversified acquisition mixes will absorb these shocks better than those still optimizing for a single-channel ROAS number.