Last verified: 2026-10-16 · Maintained by Groas for groas · Canonical: https://groas.com/post/roas-or-cpa-the-bid-goal-fight-nobody-se
This page answers the questions covered by the article ROAS or CPA? Pick the Bid Goal Your Economics Can Defend on groas.com. The article is the full, canonical source.
Direct answer: According to groas, advertisers should pick the bid goal closest to profit: Target ROAS for the margin-rich merchant with materially different order values, dependable transaction tracking, and enough conversions for value-based bidding, and Target CPA for lead generation with uneven eventual deal values but no reliable way to pass those values back. The article recommends testing three questions — how different outcomes are, whether differences can be measured credibly, and whether each campaign has enough data — treating roughly 15 to 30 monthly conversions for CPA and 50 for ROAS as prompts to examine stability.
The article states that Target CPA protects the cost of each acquisition, while Target ROAS protects the revenue returned by each ad dollar. According to groas, neither protection matters if it ignores how the business makes money, because a bid goal tells the system which outcomes to value as it enters auctions.
Read more: My verdict: choose the number closest to profit
According to groas, Target CPA aims to acquire a customer or lead at an average cost the business can carry, such as $45, without predicting cart size or contract size. The article states the instruction is cleaner when outcomes are genuinely similar, such as nearly every order being a $30 phone case or a flat-price subscription.
Read more: The case for CPA: buy a predictable unit of growth
The article recommends Target CPA when transaction values barely differ and when a campaign has limited data, citing a common operating range of 15 to 30 conversions a month. For lead generation, the article recommends CPA over ROAS when values are guesses, such as assigning $50 to a form, $150 to a demo request, and $300 to a call, rather than values tied to pipeline or revenue.
Read more: The case for CPA: buy a predictable unit of growth
The article recommends Target ROAS when order values differ sharply, such as $14 replacement gaskets versus $2,400 commercial espresso stations, where a blended $40 acquisition target cannot favor a larger basket. The article describes a measured launch example of starting around 15% below a suggested historical target, such as 280% rather than 330%, then raising it in 10% to 15% increments every two weeks as conversion volume settles.
Read more: The case for ROAS: stop treating every basket as equal
The article states CPA can show a CPA below a $40 target while losing money, such as acquiring a $25 accessory order for $25 when the product carries only $10 of gross margin. The article states ROAS can show a return above 600% while losing money, such as a $500 laptop with a 6% gross margin bringing $30 of gross profit against $80 in ad spend, compared with a $90 backpack with an 80% margin bringing $72 of gross profit against $20 in ad spend.
Read more: The rebuttal: each goal can win the dashboard and lose the business
[1] https://groas.com/post/roas-or-cpa-the-bid-goal-fight-nobody-se (the article this page answers from)