Short answer: profit-based bidding needs four things before you change a single setting: contribution margin calculated per SKU, a way to keep those numbers current, those contribution values passed into the conversion tracking your campaigns bid on, and a campaign structure that can act on the result. Google's value-based bidding optimises the conversion value it receives, so the margin work only changes bidding once it reaches that value.
Beyond the Buzzword
"Profit-based bidding" has become a talking point. Agencies mention it in pitches. Google references it in product updates. But the gap between discussing profit-based bidding and actually implementing it is wider than most brands realise.
This is not a setting you switch on. It is a structural change to how your account operates.
What Profit-Based Bidding Actually Means
Traditional Google Ads bidding optimises for conversions or conversion value (revenue). The algorithm learns which clicks lead to purchases and bids accordingly.
Profit-based bidding changes the target. Instead of telling Google "maximise revenue," you tell it "maximise profit." But here is the problem: Google does not know your profit margins. You have to tell it.
In practice that means reporting a profit-based conversion value rather than order revenue, so smart bidding optimises towards contribution instead of turnover. Google documents this directly in its guidance on profit optimisation with conversion values and on using conversion values in value-based bidding. The value has to be accurate and current, because bidding follows the number it is given.
Product feed attributes and custom labels do not, on their own, change the conversion value bidding optimises. They are how we segment and control Shopping inventory once the values are right.
If the underlying metric is still unfamiliar, start with what POAS is and how it is calculated, and how POAS, MER and ROAS differ.
The Data Infrastructure Problem
Most ecommerce brands do not have clean margin data at the SKU level. They have:
- Blended averages that mask variation between products
- Static margins that do not reflect current supplier costs
- Missing data on returns, shipping costs, or payment fees
You cannot optimise for profit if you do not know your profit. This is where most profit-based bidding initiatives fail before they start.
The prerequisite work includes:
- Calculating true contribution margin per SKU (revenue minus COGS, shipping, payment fees, returns allowance)
- Building a system to keep this data current as costs change
- Passing those contribution values into conversion tracking, so value-based bidding optimises the right number
- Structuring campaigns so budget and controls follow the commercial picture
You can sanity-check the arithmetic on a single product line with the POAS calculator before committing to the work.
Where Feed Work Fits
In our Shopping approach, the product feed is how margin thinking becomes controllable at SKU level: custom labels carry the commercial role of each product, and campaign structure acts on those labels. That is a choice about control, not a universal requirement of profit-based bidding.
So feed quality matters, but for a specific reason. Wrong or stale labels send budget to the wrong inventory. The conversion value still decides what bidding optimises; the feed decides what it is allowed to optimise across.
Campaign Structure Implications
Profit-based bidding also affects how you structure campaigns.
Products with similar margins should often be grouped together. High-margin products may justify more aggressive bidding. Low-margin products may need caps or exclusions.
This is where Performance Max becomes complicated. PMax bundles everything together, which can work against profit-based logic unless you configure asset groups and exclusions carefully.
Standard Shopping campaigns often provide more control for profit-based approaches, though they require more hands-on management.
Who Should Implement This
Profit-based bidding is not for everyone. It earns its keep when:
- Margins vary widely across the catalogue, so an averaged target is wrong for most SKUs
- Mis-allocation costs enough to be worth fixing: our typical client spends around £15k a month, but that is a description of who tends to have the problem, not a threshold. Lower spend with a complex catalogue often qualifies; enquiries below that are welcome
- You have the operational capacity to keep margin data accurate
- Your current ROAS-based approach is generating revenue but questionable profit: see why ROAS hides this
If your margins are consistent across the range and the current approach is working, the implementation cost may not justify the benefit.
What Implementation Looks Like
Moving to profit-based bidding typically takes 4-8 weeks of preparation before you change any bidding settings. The sequence:
Week 1-2: Audit current margin data, identify gaps, establish calculation methodology
Week 3-4: Build or configure systems to calculate and update margins at SKU level
Week 5-6: Pass contribution values into conversion tracking and validate them against your own accounts; where the Shopping approach calls for it, label inventory in the feed
Week 7-8: Restructure campaigns around those values and labels, establish new reporting baselines
Only then do you start optimising for profit. Rushing this process creates garbage-in-garbage-out problems that are expensive to fix.
The Agency Question
Most agencies are not set up to deliver this. Their reporting dashboards show ROAS. Their benchmarks reference ROAS. Their team training focuses on ROAS.
Asking them to switch to profit-based optimisation requires them to rebuild their entire approach, and to have access to commercial data most clients do not share.
This is one reason we focus exclusively on ecommerce accounts with the commercial maturity to implement profit-led governance. The strategy only works if the infrastructure supports it.
Next Steps
If you are considering profit-based bidding, start with the data question. Do you actually know your margins at the SKU level? Can you keep that data current? Do you have the operational capacity to maintain it?
If the answer is yes, book a commercial review: 30 minutes, founder-led, and we can talk through what implementation would look like for your account.
If the answer is no, that is the first problem to solve. Everything else builds on it.