Across consumer ecommerce benchmarks published in 2026, median Google Shopping CPCs land between $0.89 and $1.32, with ROAS figures ranging from 2.81 to 4.28 depending on the dataset. Those numbers come from DTC and retail accounts. If you're running a B2B catalog with minimum order quantities, quote-assisted SKUs, and margin spreads that vary by a factor of five across product lines, those benchmarks are decoration. They don't tell you what a qualified industrial lead or a $14,000 reorder is worth inside your pipeline.

That gap between consumer Shopping benchmarks and B2B commercial reality is where most Shopping campaigns quietly bleed budget.

Structure Comes Before Bidding

Google Shopping matching is driven by Merchant Center product data, not keywords. Standard Shopping uses product groups to organize inventory, and each group can receive a different bid. The instinct is to group by category and move on. The problem: two SKUs in the same category can have completely different margin profiles and conversion paths. A high-revenue, low-margin commodity fastener and a high-margin maintenance kit both live in "industrial supplies." Holding them to the same ROAS target either starves the kit of budget or lets the fastener eat spend it can't justify on contribution margin.

Google allows up to five custom labels in Shopping feeds (custom_label_0 through custom_label_4). Use them to encode margin tier, inventory risk, product priority, and purchase path (direct checkout vs. quote request). Don't put raw cost data in a public-facing feed; controlled labels and value rules let you steer decisions without exposing margin details.

Not a template to copy blindly. Your label taxonomy should map to your unit economics and pipeline stages, not to Google's default product type hierarchy.

B2B Compliance Is a Feed Problem

Google's Merchant Center guidance is explicit: Shopping offers should reflect prices, shipping, and checkout terms that work for the broad Shopping audience, including individual buyers. You can't gate checkout behind a business account and expect the listing to stay approved.

For products sold only in minimum quantities, Google's B2B guidance says to submit the total price for the minimum purchasable quantity, not an unpurchasable per-unit price. This creates a merchandising trade-off: showing the total price for a case of 48 units makes your listing look expensive next to a consumer competitor selling one. Teams have to decide whether that SKU belongs in Shopping at all, or whether it's better served by Search with a lead-gen landing page. The alternative (submitting a per-unit price nobody can actually buy) gets you disapproved. The right call depends on whether the SKU's contribution margin justifies Shopping visibility at the real, purchasable price point.

Bidding Is a Sequence, Not a Launch-Day Decision

Google has extended Smart Bidding Exploration to Shopping campaigns and introduced AI Max for Shopping, positioning AI-driven enhancements while stating advertisers retain product-targeting controls and bidding flexibility. More automation doesn't eliminate the need for structure. If your product data and segmentation are weak, automated bidding optimizes toward the wrong products or values.

  1. Build measurement first. Conversion tracking, revenue or value capture, feed eligibility, clean campaign segments. Use conservative bidding with a defined learning goal.
  2. Move to value-based bidding once purchase values and lead values are reliable. For quote-assisted products, enhanced conversions for leads and offline conversion imports link qualified opportunities back to Google Ads.
  3. Add efficiency targets last. Set Target ROAS only after value signals have stabilized. Change targets incrementally, document the rationale, and let results settle before adjusting again.

Success = stable cost per qualified opportunity within margin guardrails. Stop-loss = CPA on quote-assisted SKUs exceeding 2x your target for two consecutive weeks without corresponding pipeline movement.

Benchmarks Are Context, Not Targets

According to Top Growth Marketing's dataset across 28 DTC brands, median Shopping CPC was $1.02 with ROAS of 4.13. Olifant Digital reported a Shopping conversion rate of 3.9% and ROAS of 3.44 across nine ecommerce accounts. Both are consumer reference points. For a B2B catalog where the "conversion" might be a quote request with a 90-day close cycle, platform ROAS doesn't reflect pipeline value.

Your bidding guardrails should come from your own unit economics: contribution margin per SKU tier, average close rate on quote-assisted leads, and pipeline velocity by product category. Borrowing someone else's ROAS target is borrowing someone else's business model.

Structure is the input. Bidding is the output. Get them backwards and automation just scales the mistake.