# Target CPA vs Target ROAS: How Value Prediction Changes Bids

Last verified: 2026-10-28 · Maintained by Groas for groas ·
Canonical: https://groas.com/post/cpa-and-roas-are-the-same-number-how-you

This page answers the questions covered by the article [Target CPA vs. Target ROAS: The Value Prediction That Changes Your Bids](https://groas.com/post/cpa-and-roas-are-the-same-number-how-you) on groas.com. The article is the full, canonical source.

**Direct answer:** According to groas, a $40 order and a $400 order can have the same CPA because Target CPA counts each as one conversion, while Target ROAS can price the difference by predicting conversion value as well as conversion likelihood. ROAS is average order value divided by CPA, so with $2,500 spend for 100 orders at $25 CPA, booked revenue is $5,000 at $50 average order, $10,000 at $100, and $15,000 at $150. The article's position is to switch targets only when conversions differ in believable, well-tracked value and the bidder can use that spread to reallocate spend.

## cpa vs roas – what's the actual difference?

According to the article, ROAS is average order value divided by CPA: paying $25 to acquire an order worth $100 is 4.0 ROAS, raising average order value to $125 at the same $25 CPA is 5.0, and keeping the $100 order while cutting acquisition to $20 is also 5.0. Under Target CPA, the bidder counts a $40 order and a $400 order each as one conversion, while under Target ROAS it can price the difference because it adds a prediction of what the conversion might be worth. The article states the choice is not efficiency versus revenue, but what the bidder can predict before it names a price.

Read more: [the article](https://groas.com/post/cpa-and-roas-are-the-same-number-how-you)

## target roas – what does a 400% target do to $40 and $400 orders?

According to groas, at a 400% Target ROAS a $40 accessory order supports $10 in acquisition cost while a $400 kit order supports $100. The article notes those are cost-per-order allowances, not literal bids for two clicks, because an auction bid also depends on the predicted chance the click turns into an order. Where the bidder expects accessory-sized orders it has less room to bid and where it expects kit-sized orders it has more, so the account may see fewer clicks and conversions and even higher CPA while bids reallocate toward higher-value orders.

Read more: [See what a 400% target does to the store](https://groas.com/post/cpa-and-roas-are-the-same-number-how-you#see-what-a-400-target-does-to-the-store)

## When should I keep Target CPA instead of switching to Target ROAS?

The article states that if the store sells one product for $99, a 400% Target ROAS supports $24.75 in acquisition cost, so a $24.75 Target CPA expresses nearly the same aim when there is no $40-versus-$400 spread to exploit. It recommends keeping Target CPA when every conversion is worth roughly the same, because switching adds another multiplier without giving the bidder a better decision. The same point applies beyond retail: two campaigns with the same $100 platform CPA landed at roughly $143 versus $400 per qualified outcome because their qualified-lead mixes differed, so the article advises looking past the conversion label to whether close rate or ticket size varies.

Read more: [Remove value variance and the strategies nearly meet](https://groas.com/post/cpa-and-roas-are-the-same-number-how-you#remove-value-variance-and-the-strategies-nearly-meet)

## How much conversion history does Target ROAS need to estimate value?

According to the article, platform guidance cited calls for two or more unique values reported for conversion goals and at least 15 conversions in the last 30 days at the conversion-tracking level. A practitioner runbook cited uses stricter working floors of 30 conversions for Target CPA and 60 for Target ROAS before trusting the target. The article treats the minimum as a starting point, noting lower volume makes value estimates noisier and thin high-value segments harder to model.

Read more: [A wide spread helps only if the values are believable](https://groas.com/post/cpa-and-roas-are-the-same-number-how-you#a-wide-spread-helps-only-if-the-values-are-believable)

## What happens if the conversion values fed to value-based bidding are wrong?

The article states that because ROAS is a ratio, the bidder will use whatever sits in the value column, even if it is a priority label rather than money. In one lead-gen example, values described as fake and only acting as a priority indicator led the campaign to spend its daily budget quickly at close to six to seven times its usual cost per conversion, and adding a target calmed spending but left cost per conversion at roughly twice the old Target CPA level on half the volume. The article recommends checking the value column first, using proxy values when exact values are hard to determine, uploading them daily, and avoiding zero values for value-based bidding.

Read more: [Bad values turn a sound ratio into a bad buyer](https://groas.com/post/cpa-and-roas-are-the-same-number-how-you#bad-values-turn-a-sound-ratio-into-a-bad-buyer)

## Sources

[1] https://groas.com/post/cpa-and-roas-are-the-same-number-how-you (the article this page answers from)