Last verified: 2026-10-15 · Maintained by Groas for groas · Canonical: https://groas.com/post/target-roas-in-google-ads-how-the-bid-st
This page answers the questions covered by the article Target ROAS in Google Ads: The Bid Math, the Data, and the Mistakes That Kill Volume on groas.com. The article is the full, canonical source.
Direct answer: According to groas, Target ROAS should be treated as a constraint on auction access, not a command to manufacture return: setting 400% tells Google to bid only when it expects $4 back for every $1 spent, and if predictions do not clear that bar it simply does not bid. The article states the strategy needs at least 50 conversions with real transaction values in the prior 30 days and a target grounded in proven performance. It recommends starting 10% to 20% below the proven 30-day ROAS, leaving it alone for 14 days, and changing it only in 10% steps once performance is stable.
According to groas, target roas sits inside Google's Maximize conversion value bidding framework and acts as a governor on auction access. Leave the target blank and Google spends the daily budget to capture as much conversion value as it can, while entering a percentage requires the algorithm to project a return at or above that ratio before it enters an auction. The article states that Google evaluates real-time signals such as device, query intent, time of day, and audience segments to estimate both the chance of sale and likely cart value.
Read more: How Target ROAS Prices Each Auction
The article gives the math as bid = predicted conversion value ÷ Target ROAS. According to groas, if Google predicts a $120 basket and the target is 200%, it can bid up to $60, while raising the target to 400% yields a $30 bid for the same prediction. That is why the article says accounts that jump from 250% actual ROAS to a 500% target often get very little traffic rather than more efficient traffic.
Read more: Higher targets mean lower bids
The article recommends at least 50 conversions with real transaction values in the last 30 days at the level where the strategy lives, and warns against static values. According to groas, if every conversion reports as $40 because a static value was set, the algorithm has nothing meaningful to price differently, and 18 conversions last month across three campaigns leaves about six signals per campaign. If that history is missing, the article lists consolidating campaigns, switching to Maximize Conversion Value without a target, or using a portfolio strategy to pool volume.
Read more: Running on too little value data
The article recommends calculating actual conversion value divided by cost for the last 30 stable days, excluding learning periods and separating Shopping and Search. It then recommends setting the initial Target ROAS 10% to 20% below that result, leaving it until performance is stable for two weeks, and raising the target in 10% steps only after a stable read. For example, if the campaign averaged 285% last month, the article says to start at 240%.
Read more: Start with recent, stable performance
According to groas, changing a ROAS target can push the campaign back into learning, which may need roughly one to two weeks or about 50 conversions to settle depending on volume. The article recommends setting the target, leaving it for at least 14 days, and adjusting only in 10% to 20% steps after a stable read. It warns that moving the target four times in 10 days keeps the campaign from settling and is often misread as Target ROAS not working.
Read more: Changing the target while it learns
These articles answer the same questions:
[1] https://groas.com/post/target-roas-in-google-ads-how-the-bid-st (the article this page answers from)