“Let’s aim for 500%.”
Why? Because that was last quarter’s target? Because the platform suggested it? Because a round number feels more like a strategy?
A bidding target is a commercial instruction expressed through a measurement system. If the value entering that system is revenue, the bidder isn’t automatically accounting for your refunds, product margins or cash constraints. An impressive return on ad spend can coexist with a disappointing business.
First, check what the machine is being paid to chase
Google’s target ROAS bidding seeks conversion value against an average return target. The value you supply matters. A purchase value based on gross basket revenue gives a different instruction from one based on contribution. Neither a target nor a platform recommendation establishes that the resulting sales are incremental.
Before debating whether the target should be 400% or 500%, inspect the conversion actions used for bidding, the values attached to them and their treatment of cancellations and refunds. Check that primary outcomes and supporting actions aren’t both being counted as if they were separate sales.
The commercial question is straightforward: does a higher conversion value reliably mean a more valuable outcome for the business? If the answer is inconsistent, automating harder can scale the inconsistency.
Translate the economics before touching the target
Suppose a £100 attraction booking produces £30 of expected contribution before advertising, after supplier payments, payment fees, expected cancellations and other variable costs. Spending £30 to acquire it leaves nothing to cover fixed costs or profit. On that simplified first-order basis, the break-even revenue ROAS is £100 divided by £30, or 3.33x.
Now suppose the business needs £10 per order left after media. The allowable advertising cost becomes £20 and the required revenue ROAS becomes 5x. The target reflects the £10 requirement as well as the margin.
Change the contribution to £20 while keeping the same £10 requirement and allowable advertising cost falls to £10. The corresponding revenue ROAS rises to 10x. The same booking value can require a very different acquisition threshold. For an agent or marketplace, gross booking value and retained revenue are also different measures. Define which one enters the ROAS numerator and keep the margin denominator consistent.
These thresholds assume the order is additional and the costs are complete. A platform-reported 5x doesn’t establish either. Use the calculation to define the commercial requirement, then examine how well the bidding measure represents it.
An average target won’t protect every order
A blended target can hide expensive differences. High-margin products may subsidise low-margin products. Existing customers may make acquisition performance look healthier. A market with stronger repeat behaviour may tolerate a different first-order cost from one with weaker retention.
If values can represent those differences reliably, they may give the bidder a better instruction. If they can’t, consider whether the material differences justify separate controls. Every extra split comes with a cost in complexity and available signal; don’t build a separate campaign for every economic nuance.
Future contribution deserves particular care. Repeat purchases can support a higher acquisition cost, but forecast lifetime value isn’t cash in the bank. Use cohort evidence, a defined horizon and a downside case. Agree how long the business can fund the gap before payback.
A tighter target can make the dashboard prettier and the business smaller
Google notes that an excessively high target can restrict traffic. That creates a familiar trap: raise the target, lose volume, celebrate a higher average return and overlook the contribution left on the table.
The reverse move needs discipline too. Relaxing the target can open more volume, but the additional spend still has to earn its place. Compare the extra cost with the additional contribution you expect and subsequently observe. A higher total sales figure alone won’t settle that question.
Separate a budget constraint from an efficiency constraint. A campaign unable to spend because its target is restrictive presents a different decision from one consistently reaching its budget. Changing both at once makes the result harder to interpret.
Write the commercial instruction in plain English
Before changing a target, write down the outcome you’re funding, the value definition, allowable acquisition cost, required contribution after media, payback limit and customer mix you’re prepared to buy. Add the proposed change, expected volume response and the evidence that would trigger reversal.
Evaluate after allowing for conversion delay and enough time for the outcome you care about to emerge. Compare performance with the average target over the period if the target changed, rather than judging every historical day against today’s setting.
A useful recommendation sounds like this: test a lower target within a fixed spending limit because current contribution supports more acquisition, then review additional contribution and customer mix once conversions have matured.
“Google recommended it” is a record of where the suggestion came from. Your business still needs an investment case.
GV Marketing helps connect bidding decisions to customer economics, so the account is working towards growth the business can afford.
Sources
Read next: ROAS and profitability and brand-search incrementality.