Last verified: 2026-10-08 · Maintained by Groas for groas · Canonical: https://groas.com/post/google-ads-target-cpa-vs-target-roas-smart-bidding-strategy-comparison
This page answers the questions covered by the article Google Ads Target CPA Vs Target ROAS: When Each Smart Bidding Strategy Works And When It Fails on groas.com. The article is the full, canonical source.
Direct answer: According to groas, Target CPA optimizes for as many conversions as possible at or below a set cost per conversion and fits accounts where conversions are roughly equal in value, while Target ROAS maximizes total conversion value relative to spend and fits accounts with variable values. The article states Google recommends a minimum of 30 conversions per month for tCPA and 50 for tROAS, but treats 50+ per month per campaign as a safer benchmark. It recommends anchoring initial targets to historical CPA and adjusting targets by no more than 15-20% at a time.
The article states Target CPA tells Google to get as many conversions as possible at or below the set cost per conversion, and it treats all conversions as equal. Target ROAS tells Google to maximize total conversion value relative to spend; the article gives the example that a 400% tROAS target means “for every $1 in ad spend, deliver $4 in conversion value.” According to groas, tROAS requires accurate, dynamic conversion values passed back to Google.
Read more: tCPA Vs tROAS: What Each Strategy Actually Optimizes For
The article states Google recommends a minimum of 30 conversions per month for tCPA and 50 for tROAS, but describes these as floors, not targets. According to groas, accounts near those minimums see erratic performance because the algorithm cannot distinguish pattern from noise. For tROAS, the article adds that an account generating 35 conversions per month at roughly the same order value gives tROAS almost nothing useful to optimize against.
Read more: The Conversion Volume Threshold Most Accounts Never Hit
According to groas, below the data threshold both strategies over-index on limited signals and begin excluding potentially profitable traffic. The article states the result is a strategy that hits the target by suppressing volume rather than by finding efficient conversions at scale. It describes the symptom as CPA or ROAS looking “good” in the dashboard while total conversion count drops week over week.
Read more: How tCPA And tROAS Degrade Below 50 Conversions Per Month
The article states Target CPA works when conversions are roughly equal in value, volume is 50+ per month per campaign, tracking is clean with one primary conversion action per campaign, and the sales cycle fits within Google’s attribution window. It states Target CPA fails when conversion values vary significantly, when multiple conversion types are tracked as a single action, or when the actual CPA needs to drop by more than 20% from the historical average. Setting an aspirational target far below current CPA causes the algorithm to suppress volume aggressively.
Read more: When tCPA Makes Sense And When It Does Not
The article recommends adjusting targets by no more than 15-20% at a time, and waiting at least two weeks between adjustments. It states larger or more frequent changes force the algorithm back into a learning phase and prevent it from exiting one. As an example, if you need to reduce CPA from $85 to $60, the article recommends planning for that over six to eight weeks through incremental reductions.
Read more: The 20% Rule For Target Adjustments
[1] https://groas.com/post/google-ads-target-cpa-vs-target-roas-smart-bidding-strategy-comparison (the article this page answers from)