Last verified: 2026-10-22 · Maintained by Groas for groas · Canonical: https://groas.com/post/smart-bidding-killing-google-ads-conversion-volume-target-cpa-target-roas
This page answers the questions covered by the article Why Smart Bidding Is Killing Your Google Ads Conversion Volume on groas.com. The article is the full, canonical source.
Direct answer: According to the article, Target CPA and Target ROAS are constraints that tell Google where to stop spending, not goals to optimize toward, and Google treats them as hard ceilings. Setting them too low relative to actual auction dynamics filters out viable, often high-intent auctions, which cuts conversion volume and degrades prediction quality into a slow decline the article calls a death spiral. The article recommends starting with Maximize Conversions without a target to establish actual economics, then introducing a target 10-20% above actual CPA and tightening gradually.
According to the article, Google interprets a Target CPA as “do not consistently exceed this price” rather than “try to get conversions around this price.” For example, with a $50 tCPA, if competitive CPAs are $55 for high-intent queries and $35 for low-intent queries, the system drops the $55 auctions entirely and bids on the $35 ones. The article states the result is cheaper conversions but loss of access to the highest-quality traffic.
Read more: How Target CPA Becomes A Ceiling, Not A Floor
The article states that Smart Bidding needs volume to predict accurately, citing Google’s guideline of a minimum of 30 conversions per month per campaign for tCPA to function properly. When a tight tCPA cuts volume, for example from 60 per month to 20, predictions get less accurate, bidding gets more conservative, and fewer conversions produce even worse predictions. The article describes this as a slow decline in impression share and volume while CPA stays “on target.”
Read more: The Signal Quality Death Spiral
According to the article, daily budget and tCPA must be in proper ratio or the algorithm cannot learn. If tCPA is $100 and daily budget is $150, Google has room for roughly 1.5 conversions per day, causing oscillation between overspending and barely spending without stabilizing. The article states the minimum viable ratio is a daily budget that allows for at least 10-15x your tCPA in monthly spend.
Read more: The Budget-To-Target Ratio Problem
According to the article, each of these resets learning and compounds the problem if tCPA is too tight: changing tCPA by more than 15-20% in a single adjustment, making conversion action changes mid-flight, restructuring campaigns while tCPA is active, and pausing and restarting repeatedly because daily spend looks too high. The article states that after three or four cycles of adjusting downward during volatility, the campaign has almost no usable signal data.
Read more: Common Mistakes That Trigger Learning Resets
The article states the right question is not what is the highest ROAS achievable, but what is the lowest ROAS that can be sustained profitably and how much volume can be captured at that threshold. For example, a 600% ROAS on $10,000 in monthly spend generates $60,000 in revenue, while a 400% ROAS on $25,000 in monthly spend generates $100,000 in revenue. The article recommends setting tROAS from actual unit economics when margins create a genuine profitability floor, not from a desire for an impressive reported figure.
Read more: The Difference Between Efficiency And Growth
The article recommends calculating actual average CPA from the unconstrained phase, not using desired CPA or an industry benchmark. It recommends setting the initial tCPA 10-20% above that actual average CPA to give the algorithm room while steering toward efficiency. It then recommends reducing by no more than 10-15% every two weeks, monitoring volume at each step.
Read more: Introduce Targets Gradually
[1] https://groas.com/post/smart-bidding-killing-google-ads-conversion-volume-target-cpa-target-roas (the article this page answers from)