A 2.81x ROAS across 23 ecommerce accounts sounds healthy. For a B2B catalog where gross margins range from 15% on commodity parts to 60% on proprietary accessories, that number is almost meaningless. A brand hitting 2.81x on a 40% blended margin is barely breaking even before ad spend; on a 20% margin SKU, it's losing money on every order. The gap between "Google says we're profitable" and "finance agrees" is where most B2B Shopping budgets die.

Why blended CPC benchmarks mislead B2B teams

Published Shopping benchmarks are category-level ecommerce proxies, not B2B-specific data. Review-Collect reports an average Shopping CPC of $0.66 and a conversion rate of 1.91%, but Chemical & Industrial converts at just 0.83%, while Medical Supplies converts at 2.94% on a $0.63 CPC. Office & Business Needs sits at $1.09 CPC with a 2.45% conversion rate. These are directional, not predictive for your account.

The Snow Agency's ecommerce benchmark study found a Shopping CPC of $1.32, a 2.24% conversion rate, and a 2.81x ROAS across 23 accounts. Useful as a sanity check. Dangerous as a planning target. Your SKU mix, competitive set, and B2B purchase path (quotes, POs, sales-assisted orders) will produce numbers that look nothing like these aggregates. Anchor your budget to your own unit economics, not a benchmark.

The metric finance actually wants: contribution profit per click

Revenue ROAS can look healthy while profit after ads is weak. A $200 order at 2.5x ROAS costs $80 in ad spend. If the product's gross margin is 30% ($60), you're underwater before pick-and-pack. The break-even ROAS formula (1 divided by gross margin rate) tells you a 40% margin product needs 2.5x just to cover cost of goods, with nothing left for fulfillment, returns, or the ad itself.

The fix: build a contribution model outside of Google Ads. Revenue minus COGS, discounts, payment fees, shipping subsidy, pick-and-pack, returns allowance, and channel fees gives you contribution per order. From there, calculate an allowable acquisition cost per margin tier and a break-even CPC that actually means something. Google's Smart Bidding optimizes toward the conversion values you report. Feed it raw revenue, it chases revenue. Feed it contribution-aware values (or a defensible proxy), and the algorithm's incentives start aligning with finance's definition of success.

The operational catch

Passing true profit-based values into Google Ads is operationally heavy. Some teams start simpler: segment products into margin tiers using Merchant Center custom labels (high, mid, low, exclude), then set different ROAS targets by tier. Labels don't transmit margin data to the bidding algorithm directly, but they let you control spend allocation by economics rather than catalog size.

Over-segmentation creates its own problem. Too many low-volume segments starve Smart Bidding of conversion data, and performance degrades. The trade-off: you lose per-SKU precision but gain enough signal for the algorithm to learn.

How to test a budget increase without a blank check

"Limited by budget" in Google Ads tells you how much traffic you're leaving on the table with current settings. It says nothing about whether that incremental traffic will convert profitably. Treat it as a diagnostic signal, not a spending recommendation.

Setup: Pick one margin tier or product group. Document the pre-change baseline (spend, CPA, contribution per order, conversion volume). Set a maximum incremental spend (e.g., +25% of current daily budget for that group). Define a 2-4 week test window.

Hypothesis: If we increase daily budget by 25% on high-margin SKU group A, then incremental contribution per order will remain above $X because the lost impression share data suggests qualified demand exists at current CPCs.

Success metric: Incremental contribution margin on the added spend (not blended with historical performance). Guardrail: Blended CPA for the group doesn't rise more than 15%. Stop-loss: If incremental ROAS drops below break-even for that tier's margin, pause the increase and revert.

Assess marginal results on the added spend itself. Blending before-and-after numbers lets strong historical performance mask declining returns on the new dollars.

The offline conversion gap most B2B teams ignore

For B2B purchase paths that include quotes, purchase orders, and sales-rep-assisted orders, optimizing to online "purchase" alone undercounts value. A buyer clicks a Shopping ad, submits an RFQ, and closes a $30,000 order through a sales rep eight weeks later. If you're only measuring the form fill, Google's bidding algorithm never learns which clicks produce real revenue.

The measurement design that survives scrutiny: track the most meaningful measurable stage (qualified opportunity, closed-won), upload offline outcomes with values aligned to the business goal, and reconcile back to ad spend. Without it, Target ROAS is optimizing toward a proxy that may have little relationship to actual pipeline.

What a defensible Shopping budget actually looks like

It ties every incremental dollar to a specific product group, a defined time window, an expected contribution outcome, and a rollback condition. It separates proven earners from test cohorts from strategic bets, with different success thresholds for each. It acknowledges that published benchmarks are someone else's averages and replaces them with first-party data as fast as possible.

The CFO doesn't need to see a 2.81x ROAS slide. They need to see which SKUs earn their ad spend back after all variable costs, which ones are being tested with a capped budget and a kill switch, and what happens if the answer is "pause." That's the difference between a budget request and a budget defense.