Last updated: February 10, 2026
Switching to Target ROAS bidding is one of the highest-leverage moves you can make in Google Ads. It's also one of the easiest ways to tank a perfectly good campaign overnight.
Every week, advertisers make the jump from manual bidding or Maximize Conversions to Target ROAS, pick a number that seems reasonable, and wake up the next morning to a campaign that's either stopped spending entirely or started haemorrhaging money. The "limited by budget" warning appears out of nowhere. Conversions drop to zero. Or worse, conversions keep flowing but at a return that's nowhere near profitable.
The problem isn't Target ROAS itself. It's that most advertisers set the wrong target, set it at the wrong time, or change it in ways that send Google's algorithm into a tailspin. And the guides that exist on this topic either skip the actual math entirely or treat it like a simple settings toggle rather than the strategic decision it actually is.
This guide covers everything you need to get Target ROAS right: how to calculate your target from your real profit margins, why certain margin ranges produce specific ROAS targets, what causes the dreaded "limited by budget" problem after switching, the mistakes that reset Google's learning phase and cost you weeks of poor performance, and when to stop fighting with manual targets altogether and let an autonomous system handle it dynamically.
Before you type a single number into Google Ads, you need to understand the relationship between your profit margins and the ROAS required to maintain them. This is the step most advertisers get wrong, and it's the reason so many campaigns either lose money or strangle their own volume.
Your break-even ROAS is the minimum return you need just to cover your costs. Below this number, every dollar spent on ads loses money. The formula is straightforward.
Break-even ROAS = 1 / Gross Profit Margin (as a decimal)
If your gross margin is 50%, your break-even ROAS is 1 / 0.50 = 2.0 (or 200%). If your margin is 25%, your break-even is 1 / 0.25 = 4.0 (or 400%). If your margin is 40%, it's 1 / 0.40 = 2.5 (or 250%).
This is the number below which you are literally paying Google to lose money. And yet, a shocking number of advertisers have never calculated it.
This is where precision matters. Your gross profit margin for this calculation should include the cost of goods sold (COGS) and any variable costs directly tied to fulfilling the sale. That means product cost, shipping, packaging, payment processing fees, and any per-order costs. It should not include fixed overhead like rent, salaries, or software subscriptions, because those costs exist whether you make the sale or not.
Let's walk through a concrete example. You sell a product for $100. Your costs break down like this: product cost is $35, shipping is $8, packaging is $2, and payment processing is $3. Your total variable costs are $48, leaving a gross profit of $52 per unit. Your gross margin is 52%.
Your break-even ROAS is 1 / 0.52 = 1.92 (or 192%). Any ROAS above 192% keeps you in the black on a per-order basis.
Break-even ROAS keeps the lights on. It doesn't build the business. Your actual Target ROAS needs to be high enough that after covering COGS and ad spend, there's meaningful profit left over.
The formula for your Target ROAS with a specific profit goal is:
Target ROAS = 1 / (Gross Margin - Desired Profit Margin as % of Revenue)
Let's continue the example. Your product sells for $100 with a 52% gross margin. You want to keep 20% of revenue as profit after ad costs. Your target is 1 / (0.52 - 0.20) = 1 / 0.32 = 3.125 (or 312.5%).
Rounding up, you'd set a Target ROAS of approximately 310% to 320% in Google Ads.
Here's what that means in plain terms: for every $1 you spend on ads, Google needs to generate $3.13 in revenue. From that $3.13, you keep $1.63 in gross profit (52% margin), subtract the $1.00 ad cost, and you're left with $0.63 in profit, which represents roughly 20% of the revenue generated.
This is the most commonly searched margin range when it comes to Target ROAS, and for good reason. A huge number of ecommerce businesses and product-based companies operate with gross margins in the 40% to 50% range. Here's exactly how the math works at various profit targets across this margin range.
Your break-even ROAS is 1 / 0.40 = 250%. Below 250%, you're losing money on every sale.
If you want 10% of revenue as profit after ad costs, your Target ROAS is 1 / (0.40 - 0.10) = 1 / 0.30 = 333%. For every $1 in ad spend, you need $3.33 in revenue. From $3.33 revenue, you keep $1.33 gross profit (40%), subtract $1.00 ad cost, leaving $0.33 profit.
If you want 15% profit, your Target ROAS is 1 / (0.40 - 0.15) = 1 / 0.25 = 400%. This is the classic "4x ROAS" target that many advertisers default to. Now you understand why: it works for businesses with around 40% margins who want to keep 15% of revenue as profit.
If you want 20% profit, your Target ROAS is 1 / (0.40 - 0.20) = 1 / 0.20 = 500%. At this level, you're asking Google to return $5 for every $1 spent. That's achievable in many verticals but will significantly limit your volume.
Your break-even ROAS is 1 / 0.50 = 200%. You have more breathing room than the 40% margin scenario.
At 10% profit target, your ROAS is 1 / (0.50 - 0.10) = 1 / 0.40 = 250%. At 15% profit, it's 1 / (0.50 - 0.15) = 1 / 0.35 = 286%. At 20% profit, it's 1 / (0.50 - 0.20) = 1 / 0.30 = 333%. At 25% profit, it's 1 / (0.50 - 0.25) = 1 / 0.25 = 400%.
Notice how 50% margins give you a 400% ROAS target at 25% profit, while 40% margins produce a 400% target at only 15% profit. Same ROAS number, very different profitability. This is exactly why blindly adopting "industry standard" ROAS targets without understanding your own margins is a recipe for either wasted spend or strangled growth.
Break-even ROAS is 1 / 0.45 = 222%. At 10% profit, target is 286%. At 15% profit, target is 333%. At 20% profit, target is 400%. At 25% profit, target is 500%.
The key takeaway from all these calculations is that your ideal Target ROAS is unique to your business. Two companies in the same industry selling similar products can have wildly different optimal targets based on their cost structures. Anyone telling you to "just set it at 400%" without asking about your margins is giving you advice that has a coin-flip chance of being wrong.
This is the single most common problem advertisers face after switching to Target ROAS, and it catches thousands of people off guard every month. You had a campaign spending its full daily budget with Maximize Conversions or Manual CPC. You switch to Target ROAS. Within 24 to 48 hours, your campaign shows "limited by budget" even though nothing about your budget changed. Spend drops by 40%, 60%, sometimes 80%.
Here's what's actually happening, and it's not a bug. It's the bidding strategy working exactly as designed.
When you use Maximize Conversions, Google's goal is simple: spend your budget and get as many conversions as possible. It will bid aggressively to use every dollar. It doesn't care whether those conversions are profitable. It just wants volume.
When you switch to Target ROAS, Google's goal changes fundamentally. Now it's trying to hit a specific return on every dollar spent. It evaluates each auction individually, estimating the probability and likely value of a conversion. If Google predicts that a particular auction won't meet your ROAS target, it either bids very low or doesn't bid at all.
The result is that Google becomes selective about which auctions it enters. It stops bidding on queries, audiences, and placements where it doesn't expect to hit your target. If your ROAS target is aggressive (meaning high), Google disqualifies a large percentage of available auctions, and your budget goes unspent. That's when the "limited by budget" warning appears, because technically your budget is available, but Google can't find enough profitable auctions to spend it.
This confuses everyone the first time they see it. "Limited by budget" usually means you need more budget. But in this case, you have budget. The real constraint isn't money. It's opportunity. Google is saying: "At the return you're demanding, there aren't enough auctions worth bidding on."
This happens most often when advertisers set their Target ROAS significantly above their historical performance. If your campaigns were running at 250% ROAS under Maximize Conversions and you switch to a 500% ROAS target, you've told Google to only bid on auctions where it expects double the return you were previously getting. Of course volume collapses. You've eliminated most of the auctions your campaign was previously participating in.
The fix depends on what's causing the restriction.
If your ROAS target is too aggressive relative to your historical performance, lower it. Google's own documentation recommends setting your initial Target ROAS at or below your historical ROAS over the past 4 weeks. If your campaigns were averaging 320% ROAS before the switch, start your Target ROAS at 300% to 320%, not 500%. You can increase it gradually once the algorithm stabilises and learns.
If your daily budget is genuinely too low for your ROAS target, you may need to increase it. Target ROAS bidding works best when the daily budget is at least 50% higher than your average daily spend. This gives the algorithm room to bid aggressively in high-value auctions without hitting a budget ceiling. Remember that Google can spend up to twice your daily budget on any given day but averages out over the month. A tight budget constrains this flexibility.
If your conversion volume is too low for the algorithm to learn, that's a data problem. Target ROAS requires a minimum of 15 conversions in the past 30 days, but Google recommends 30 or more for reliable performance. With low conversion volume, the algorithm doesn't have enough signal to predict which auctions will hit your target, so it bids conservatively on everything.
This is the most common mistake and the one most directly responsible for campaigns that stop spending. When you set a Target ROAS of 800% in an account that historically runs at 350%, you're not setting an aspirational goal. You're telling Google's algorithm that 90% of available auctions aren't worth bidding on.
The algorithm will try to find the tiny subset of auctions that might deliver 800% returns. These auctions are rare, competitive, and often involve branded queries or bottom-of-funnel searches from people who were going to buy anyway. Your campaign shrinks to a trickle of highly qualified but incredibly low-volume traffic.
The fix is to start at or slightly below your historical ROAS average. If your campaigns have been running at 350% ROAS, start your Target at 330% to 350%. Once the algorithm stabilises (usually 2 to 3 weeks), increase the target by 10% to 20% increments and give each adjustment at least 2 weeks to settle before evaluating.
The opposite problem is less dramatic but equally costly. A Target ROAS of 150% when your break-even is 250% means Google will happily acquire conversions that lose you money on every sale. The algorithm will meet your target beautifully. Your bank account will not.
This mistake often happens when advertisers calculate their break-even ROAS incorrectly, use revenue instead of profit as their conversion value, or simply pick a low number to "let the algorithm learn" and forget to increase it.
Always set your Target ROAS above your calculated break-even. If your break-even is 250%, your target should be at least 280% to 300% to ensure a meaningful margin of safety.
Every time you change your Target ROAS, you reset Google's learning phase. During the learning phase, which typically lasts 1 to 2 weeks, the algorithm experiments with different bid levels and auction strategies to calibrate against your new target. Performance during this period is unreliable and often worse than your pre-change baseline.
Advertisers who panic and change their target every few days create a perpetual learning loop. The algorithm never stabilises. Performance stays erratic. And the advertiser concludes that "Target ROAS doesn't work for my business" when the real problem was that they never gave it a chance to work.
Google's own guidance is clear: after changing your Target ROAS, wait at least 2 weeks before evaluating results or making further changes. If you're uncomfortable with the performance during this period, remind yourself that 2 weeks of learning is the cost of making the algorithm significantly more effective for the months that follow.
Target ROAS requires enough historical conversion data for Google's AI to make meaningful predictions. The technical minimum is 15 conversions in the past 30 days per campaign, but real-world performance typically requires 30 to 50 conversions per month for reliable optimisation.
If you switch to Target ROAS with only 10 conversions in the past month, the algorithm is essentially guessing. It doesn't have enough data points to distinguish between a $5 click that's likely to convert at $200 and a $5 click that will bounce. The result is either overspending on low-value clicks or underspending on everything because the algorithm lacks confidence.
The best path to Target ROAS for campaigns with limited data is to start with Maximize Conversion Value (without a ROAS target) for 4 to 6 weeks. This lets the algorithm accumulate conversion value data while spending your budget. Once you have 30 or more conversions with value data, layer on a Target ROAS goal based on the actual ROAS achieved during that period.
Target ROAS optimises toward the conversion values you feed it. If those values are wrong, the algorithm will confidently optimise toward the wrong outcomes.
The most common conversion value errors include tracking revenue inclusive of sales tax and shipping (which inflates your apparent ROAS by 10% to 25%), counting duplicate conversions from the same transaction, using static placeholder values instead of dynamic revenue data, not accounting for returns and refunds (which overstate actual revenue), and having stale product prices in feed data.
Before switching to Target ROAS, audit your conversion tracking ruthlessly. Compare the conversion value reported in Google Ads against actual revenue in your payment processor or ecommerce platform over a 30-day period. If the numbers don't match within 5% to 10%, fix your tracking before touching your bidding strategy. An algorithm optimising against bad data will produce confidently bad results.
Google's Smart Bidding algorithms need time to learn whenever you make significant changes. The learning phase is the period after a change where the algorithm experiments with different bid levels, adjusts its predictions, and calibrates against your target. During this period, performance is typically worse than your pre-change average.
A new learning phase begins whenever you switch bidding strategies (for example, from Maximize Conversions to Target ROAS), change your ROAS target by more than 10% to 15%, significantly change your daily budget (increases or decreases of more than 20%), add or remove conversion actions, make major changes to campaign structure (adding or removing ad groups, keywords, or audiences), or change your geographic or audience targeting substantially.
Each of these triggers sends the algorithm back to a period of experimentation. It needs to recalibrate its models against the new parameters, and that calibration takes time and data.
During the learning phase, your bids may be higher or lower than usual as the algorithm tests different levels. Your CPC might spike 20% to 40% above normal. Conversion volume often drops as the algorithm bids conservatively in unfamiliar territory. Your actual ROAS may be significantly above or below your target.
This period typically lasts 7 to 14 days for campaigns with strong conversion volume (30 or more conversions per week). For campaigns with lower volume, learning can stretch to 3 to 4 weeks or longer because the algorithm needs more time to gather enough data points.
Every learning phase costs you money in the form of sub-optimal performance. If your campaign normally generates 100 conversions per month at a 400% ROAS, and the learning phase runs for 2 weeks at 280% ROAS, you've effectively lost roughly half a month of optimised performance.
Advertisers who change their Target ROAS every week trigger 52 learning phases per year. That means their campaign spends more time learning than performing. Compare that to an advertiser who sets a target, waits a month, evaluates, and makes one careful adjustment, experiencing perhaps 6 to 8 learning phases per year.
This is one of the most compelling arguments for automated management. Every manual adjustment to a Target ROAS triggers a learning reset that costs you time and money. An autonomous system that continuously optimises without needing to reset the algorithm's learning avoids this tax entirely.
Everything we've covered so far assumes you're manually setting and adjusting a Target ROAS. And the truth is, doing this well is genuinely hard.
You need to calculate the right target from your margins. You need to set it correctly based on historical data. You need to avoid changing it too often or too dramatically. You need to wait through learning phases without panicking. You need to monitor whether your conversion values are accurate. You need to account for seasonal fluctuations, competitor behaviour, and shifting auction dynamics. And you need to do all of this continuously, month after month, without making the mistakes that send performance off a cliff.
This is where the conversation shifts from "how to set Target ROAS" to "should you be setting it manually at all."
A static ROAS target is a fixed number applied to a dynamic marketplace. Your competitors' bids change hourly. Consumer search behaviour shifts by day of week, time of day, device, location, and dozens of other variables. Your own conversion rates fluctuate with landing page performance, seasonal demand, inventory levels, and promotional activity.
A fixed 400% ROAS target treats Tuesday morning and Saturday night identically. It treats January's competitive landscape the same as Black Friday. It doesn't know that your margins improved this month because you renegotiated supplier costs. It doesn't know that a competitor just pulled their budget, creating a temporary window of cheaper auctions you should be capitalising on.
Human managers try to account for this by periodically adjusting their targets. But "periodically" means weekly or monthly at best. And every adjustment triggers a learning phase that costs performance. It's a losing game of whack-a-mole: by the time you identify a change and adjust your target, the conditions have already shifted again.
This is precisely the problem groas was built to solve. Instead of setting a static ROAS target and hoping the market cooperates, groas continuously evaluates real-time auction conditions, conversion probability, margin data, and performance signals to make dynamic bid decisions thousands of times per day.
There's no static target to set incorrectly. There's no learning phase to trigger with manual changes. There's no gap between identifying a market shift and responding to it. The system processes information and adjusts at machine speed rather than human speed.
groas is particularly effective here because of its deep integration with Google's bidding infrastructure. Rather than working against Google's AI, groas works with it, feeding better signals and more precise optimisation goals into the same auction-time bidding system that Target ROAS uses. The difference is that the targets and signals are continuously calibrated rather than manually fixed.
The practical impact is significant. Campaigns managed by groas avoid the learning phase penalties that manual target changes create. They don't suffer from the "limited by budget" problem caused by overly aggressive static targets. And they don't leave money on the table during periods when a lower ROAS target would be profitable because market conditions temporarily favour aggressive spending.
For advertisers who've spent months wrestling with Target ROAS settings, tweaking numbers, waiting through learning phases, and trying to figure out why their campaigns keep oscillating between not spending and overspending, the shift to autonomous management isn't a luxury. It's an escape from a fundamentally unwinnable manual process.
If you're going to set Target ROAS manually, here's the sequence that gives you the best chance of success. Follow each step carefully, and resist the temptation to skip ahead.
Step 1: Audit your conversion tracking. Verify that the conversion values in Google Ads match your actual revenue within 5% to 10%. Check for duplicate conversions, ensure tax and shipping aren't inflating values, and confirm that all your conversion actions are correctly configured. Do not proceed until your data is clean.
Step 2: Run Maximize Conversion Value for 4 to 6 weeks. If you're coming from manual bidding or Maximize Conversions, don't jump straight to Target ROAS. Run Maximize Conversion Value without a target first. This lets Google's algorithm learn which auctions generate the most valuable conversions using your budget. You need at least 30 conversions with value data during this period.
Step 3: Calculate your actual ROAS from the past 30 days. Once you have 4 to 6 weeks of data under Maximize Conversion Value, look at your actual Conv. value/cost metric in Google Ads. This is your real-world ROAS baseline.
Step 4: Calculate your break-even and target ROAS. Use the formulas from earlier in this article. Determine your break-even ROAS (1 / gross margin) and your desired profit target ROAS (1 / (gross margin - desired profit %)). Compare both against your actual ROAS from Step 3.
Step 5: Set your initial Target ROAS at or slightly below your actual ROAS. Google explicitly recommends this. If your campaigns averaged 380% ROAS in the data period, set your target at 350% to 380%. Setting it higher than your actual performance is the number one cause of campaigns stopping delivery.
Step 6: Wait 2 full weeks without touching anything. No budget changes. No target adjustments. No campaign restructuring. Let the algorithm learn. Performance during this period will likely be erratic. That's normal.
Step 7: Evaluate and adjust incrementally. After the learning phase, compare your ROAS against your target and your profitability goals. If the algorithm is hitting your target and you want higher returns, increase the target by no more than 10% to 20%. Wait another 2 weeks. Repeat.
Step 8: Monitor continuously. Check your search term reports weekly for irrelevant traffic. Verify conversion values monthly against actual revenue. Review the budget status for "limited by budget" signals. Adjust your target no more than once per month unless something is genuinely broken.
Use the formula: Target ROAS = 1 / (Gross Margin - Desired Profit Margin), with both margins expressed as decimals. For example, if your gross margin is 45% and you want 15% profit after ad costs, your Target ROAS = 1 / (0.45 - 0.15) = 1 / 0.30 = 333%. Your break-even ROAS (the absolute minimum) is simply 1 / Gross Margin. At 45% margins, that's 222%.
At 40% gross margins, your break-even ROAS is 250%. For 10% profit, target 333%. For 15% profit, target 400%. For 20% profit, target 500%. The higher your profit goal, the more restrictive your target becomes and the less volume Google will deliver. Start closer to your break-even and increase gradually based on actual campaign performance.
At 50% gross margins, your break-even ROAS is 200%. For 10% profit, target 250%. For 15% profit, target 286%. For 20% profit, target 333%. For 25% profit, target 400%. Higher margins give you significantly more flexibility to set competitive targets that maintain volume while remaining profitable.
When you switch to Target ROAS, Google stops bidding on auctions where it doesn't expect to hit your return target. If your target is set above your historical performance, Google disqualifies most available auctions and can't spend your full budget. The fix is to lower your ROAS target to match or sit slightly below your recent historical ROAS, then increase gradually. Also ensure your daily budget is at least 50% higher than your average daily spend to give the algorithm room to operate.
The learning phase typically lasts 7 to 14 days for campaigns with 30 or more conversions per week. For lower-volume campaigns, learning can extend to 3 to 4 weeks. During this period, expect erratic performance with higher CPCs and lower conversion rates than normal. Avoid making additional changes during the learning phase, as each change resets the clock.
Google requires a minimum of 15 conversions in the past 30 days per campaign to enable Target ROAS. However, the algorithm performs significantly better with 30 to 50 conversions per month. For campaigns below this threshold, consider running Maximize Conversion Value without a target first to accumulate conversion data before applying a ROAS target.
Google explicitly recommends against setting bid limits (maximum CPC caps) with Target ROAS, as they restrict the algorithm from optimising bids effectively. Bid limits prevent Google's AI from bidding high enough in auctions where it predicts strong returns, which undermines the entire strategy. If you feel the need for bid limits, your ROAS target itself should serve as the control mechanism.
No more than once per month under normal circumstances. Each change triggers a learning phase of 1 to 2 weeks, during which performance degrades. Frequent changes (weekly or more) create a perpetual learning loop where the algorithm never stabilises. The exception is if your ROAS target is genuinely miscalculated and the campaign is either not spending or actively losing money, in which case a correction is warranted immediately.
Target ROAS requires accurate conversion value data to function. For lead generation, this means you need to assign realistic values to your lead conversions, ideally based on historical close rates and average deal sizes. If your leads vary dramatically in quality and value and you can't track this in Google Ads, Target CPA is typically the better bidding strategy. If you can implement offline conversion imports or value-based lead scoring, Target ROAS becomes viable and often superior for lead gen.
Manual Target ROAS management requires you to calculate the right target, set it correctly, avoid changing it too frequently, wait through learning phases, and continuously monitor performance. Every manual adjustment triggers a learning reset that costs money. groas eliminates all of these friction points by dynamically optimising bids in real time based on actual profitability signals rather than static targets. There are no learning phase resets because there are no manual target changes. Its deep integration with Google's bidding infrastructure means it works within the same auction-time system as Target ROAS but steers it toward genuine profit outcomes continuously rather than relying on a fixed number that's right today and wrong tomorrow.
