How a ROAS goal works
A ROAS goal is the value-based cousin of a cost target. Rather than naming what a result may cost, you name the return you want on the money: how much purchase value should come back for each unit of spend. The system then bids to hold that ratio across the campaign as a whole while it spends the budget, paying up for shoppers it expects to spend heavily and withdrawing where the return looks thin.
It depends completely on values reaching the platform. Purchase events have to carry an amount and a currency, from the pixel, the conversions API or a catalogue. It is also an average, not a floor — some orders will return far more and some far less, and the goal describes where the middle should sit. The old name, minimum ROAS, wrongly suggested a guaranteed lower bound.
Why a ROAS goal matters
It puts profitability rather than activity into the bidding. A shop can hold a fine cost per purchase and still lose money if the orders are small, whereas a return target ties spending directly to what comes back. For businesses with a known gross margin, it is the setting that most closely reflects the real constraint.
It also forces an honest trade-off into the open. Asking for a higher return means the system must refuse more auctions, so volume falls as the target rises. Seeing that trade-off is useful: it tells you what growth costs in margin, and whether the account has room to scale at all.
Where ROAS goals go wrong
The first trap is the number the platform reports back. Reported return is built on the platform’s own attribution and on the values your site sends, so duplicated events, delivery charges counted as revenue or a currency mismatch will all flatter it. Chasing a target against a flattered figure means optimising towards a number that does not exist in your bank account — which is why platform-reported return should always be reconciled against your actual sales.
The second is setting the target from the return you wish you had. A goal far above anything the account has achieved simply stops delivery, and the campaign underspends while looking blameless. The third is forgetting the margin question: optimising towards revenue pushes the system towards your dearest products, which are not always the ones you earn most on. In markets where cash on delivery is normal, recorded revenue and collected revenue are not the same thing, and returned parcels quietly widen the gap.
How to act on it
Run an unconstrained value strategy such as highest value first, so you know what return the account genuinely produces. Check your events for duplicates and confirm the currency. Then set the goal near the achieved figure rather than the ambition, and tighten it gradually while watching what happens to spend.
Judge the outcome on total profit, not on the ratio alone: a high return on a tiny budget can be worth less than a modest return at scale. If delivery collapses when you tighten, that is the auction telling you the margin you want is not available at that volume, and the answer lies in the offer, the products or the site rather than in the bidding rule. That reconciliation is ordinary ecommerce advertising work.