Every quarter, Google reports ad revenue growth while performance marketers watch their blended margin erode. That tension is not an accident or poor campaign setup. It is architectural.
Google's automated bidding algorithms are engineered to maximise auction liquidity and platform yield. When you set a Target ROAS or Target CPA without strict constraints, Smart Bidding treats every gross revenue dollar as equal. It does not know your product margins, shipping surcharges, or return rates unless you force it to.
If you let Smart Bidding chase raw top-line conversion value, two things happen immediately. First, the algorithm finds high-volume, low-margin transactions because they are the easiest auctions to win. Second, your ad spend inflates to hit aggregate revenue targets that quietly destroy your net operating profit.
The TL;DR: Google Ads Smart Bidding optimises for aggregate conversion value and platform auction efficiency, not your net ledger profit. To fix this, you must feed margin-weighted conversion values into tROAS, use conversion value rules, and enforce profitability guardrails before scaling spend.
Why platform yield works against advertiser margin
Smart Bidding is extraordinarily good at solving the objective function you give it. The problem is that most accounts give it the wrong objective.
When you run Target ROAS, Google calculates: Bid = Target ROAS * Expected Conversion Value
The platform has a natural incentive to maintain high auction density. If the algorithm can clear an auction by winning a conversion with a 300% gross ROAS on a product with a 10% margin, it will take the bid. On your dashboard, the campaign hits its target. In your bank account, you lost money on the sale once fulfillment and overhead are factored in.
This is why understanding how Smart Bidding reads conversions is vital. The machine views a £100 order with £5 profit exactly the same as a £100 order with £65 profit. Without custom margin weighting, the bidder naturally drifts toward the path of least resistance.
How to force Smart Bidding to bid on real profit
Fixing this requires shifting from gross revenue bidding to contribution margin bidding. Here is the operational framework I use across client accounts.
1. Shift from top-line revenue to dynamic margin feeds
Stop passing gross checkout values in your transaction tags. Instead, calculate gross margin per order before the conversion value is posted to Google Ads, or adjust values post-purchase via offline conversion adjustments.
If you run eCommerce on Shopify, WooCommerce, or custom cart infrastructure, build a margin-adjusted conversion pipe. I detail the exact setup in my guide on building a margin feed for tROAS that optimises real profit. When Google's bidding engine optimises for margin value rather than top-line revenue, it automatically deprioritises unprofitable product volume.
2. Implement conversion value rules for high-cost customer segments
Not every customer cohort carries the same fulfillment cost. Returns, shipping zones, and payment processing fees introduce hidden margin leakage.
Use Conversion Value Rules in Google Ads to down-weight:
- Geographic zones with high freight or customs costs.
- Devices with historically lower basket retention.
- Audiences with high return probabilities.
3. Enforce target ROAS floors based on blended break-evens
Most advertisers set their tROAS target based on last month's averages. That creates an anchor trap. Calculate your true contribution break-even point across ad spend, COGS, and payment fees, then add a 15-20% margin cushion as your minimum operational floor.
If your campaigns are showing strong ROAS numbers but overall profitability is stagnant, getting a Google Ads audit helps identify whether automated bidding is bleeding budget on margin-negative queries. If your account is doing this, the free audit takes 5 working days and includes a Loom walkthrough.
Case study: B2C retail account shifting to contribution bidding
A mid-market lifestyle retailer was scaling spend from £40k/month to £75k/month under standard Target ROAS (380% target). Gross reported revenue increased by 65%, but monthly net operating profit declined by 14%.
A 90-day diagnostic showed that Smart Bidding was heavily favoring low-margin promotional items and clearing auctions for repeat purchasers who would have converted organically.
Over a 90-day transition window, we implemented margin-weighted conversion tracking and tightened audience signals:
- Ad spend stabilized at £55k/month (down 26% from peak, eliminating auction waste).
- Blended net margin increased by 28-35%.
- Net profit contribution grew by 24% on 20% lower gross volume.
The account generated more actual profit with fewer total orders and significantly less customer support overhead.
Common mistakes
- Trusting dashboard ROAS blindly. A 400% ROAS on 15% margin inventory loses money after overhead.
- Letting AI Max run on default targets. Broad matching without margin constraints causes the system to grab cheap, low-value conversions across secondary networks.
- Changing targets too abruptly. Large tROAS swings (over 15-20% weekly) reset algorithmic learning phases and destabilize bidding pools.
Bottom line
- Google Ads builds algorithms to maximise auction liquidity and aggregate ad spend.
- You cannot expect an automated system to protect your profit margins unless you feed it margin data.
- Pass net contribution margin values into your conversion tags, enforce tROAS floors, and penalise high-friction conversion segments.
The advertisers thriving right now are not the ones spending the most. They are the ones who refuse to let Google treat gross revenue and net profit as interchangeable metrics.
Sources and further reading: