Target ROAS in Google Ads: Where to Start, When to Change It, and How to Tell If It Paid Off

We follow this process for every target change on every account we manage. You need the Google Ads UI, a spreadsheet, and the store's margin, and you can do every step by hand. The video above walks through one real session on a real store's account. We masked the industry and names and converted the dollar amounts, but every percentage is exactly what the data showed.

Step 1: Make sure the numbers mean what you think

Google stores a 500% target as 5.0, and reports show return as a percentage. If you line up 5.5 against a 470% result without converting, the campaign looks ready for a higher target when it's actually below the one it has. Convert both to percentages first.

Then look at the conversion value itself, because that's what the target learns from. Counting add-to-cart as a primary conversion, or giving micro-conversions a value, inflates ROAS and teaches bidding the wrong lesson. Canceled orders and financing that falls through should be sent back as negative transactions. And once a consent banner goes live, some conversions drop out of measurement, so reported ROAS falls while real sales hold steady. Factor that in before you decide a campaign got worse.

Our working table holds enabled campaigns over the last 30 days. We read each target from wherever it really lives, the campaign or a shared portfolio strategy, and next to any campaign that isn't serving we note why. A recommendation for a suspended campaign won't change a thing until it serves again.

Step 2: Set the first target from what the campaign already does

A new Performance Max campaign launches on Maximize conversion value with no target, even on an account with years of history. We add a target once that specific campaign has around 50 conversions. At that point we read its actual return and set the target right there, or about 10% below. Read that return by click date, which is how the standard columns in the Google Ads UI and the bidding itself credit value. When people take a few days to buy, ROAS by conversion date can come out noticeably different. Shopping and Search campaigns follow the same pattern: they build conversion volume without a target first, then get one at the return they actually deliver.

The classic mistake is typing in the business goal on day one. A campaign returning 300% that suddenly gets a 600% target simply stops spending. You work up to the business goal in steps, covered below.

Step 3: Read the warning signs

A target that's too high shows up as unspent budget and a campaign that only picks off the occasional high-return sale. The fix feels backward: drop the target 10-20% below the campaign's real ROAS. With fewer than 50 conversions, remove it for a week or two, see what the campaign actually delivers, and reset it there.

A target that's too low shows up as a campaign that spends every dollar while profit shrinks. A target far below what the campaign achieves behaves like Maximize conversion value with no target at all, and that mode ignores margin: it chases the priciest products first, then slides down to the thinnest returns. The one exception is a target deliberately parked at the break-even point as a safety net, so spend pulls back on its own if the market turns.

Lost impression share helps you tell the cases apart. Share lost to budget means the money ran out. Share lost to rank means the campaign is losing auctions, often because the target won't let it bid higher. Performance Max barely reports share lost to budget, so for PMax we compare average daily spend with the daily budget. When a campaign with a target spends its whole budget and still misses its return, the lever is the target, and cutting the budget only routes around the problem. Since August 17, 2026, a campaign with a target and a "Limited by budget" status steers to the target you set. If it used to beat that target comfortably, it will drift down to it, and nothing will flag it.

One more thing people miss: bidding doesn't try to beat your target. It aims for it and usually lands a little under. If a campaign is beating its target and making money, give it more budget.

Step 4: Match the answer to the campaign's job

A single fix for the whole account is almost always wrong. In the video, six campaigns got three different answers. The two product campaigns move up one step. The test campaign waits a full conversion cycle because its conversion goals were only fixed a month earlier. The brand campaign spends its full budget and stays exactly as it is.

Brand protects sales the store already owns. More money there mostly buys searches you'd win anyway, so we don't put a ROAS target on brand and look for growth in the product campaigns instead. Since one target covers everything inside a campaign, products with very different margins belong in separate campaigns. One might justify 1000% and another 500%, and each target fits its own margins. The other route is to report profit to Google instead of revenue, meaning order value minus the cost of goods. The target then works off margin, an approach usually called POAS, and one target can cover products with different margins. It means sending margin data to Google and keeping it protected, so only a small share of accounts go this way.

Step 5: Run the numbers before you lower the target

A lower target opens up new traffic and also lowers the return on all the sales you were already making. Before any cut, we figure out how much spend has to grow just to keep profit flat. You need two formulas and the margin:

• break-even ROAS = 100 ÷ margin percentage (a 25% margin means 400%),
• profit after ad spend = conversion value × margin − ad cost.

Here's the example from the video. Blended margin is 32.5%, and the target goes from 600% to 500%. At 600%, every $100 in spend leaves about $95 of profit after ad spend. At 500%, it leaves $62.50. Spend has to grow by roughly half just to break even on the change. We write that number down beforehand, because afterward it's tempting to call higher spend a win by itself.

ROAS dropping while profit climbs is the normal price of scale: double the volume and return usually dips while total profit grows. ROAS dropping while profit falls means it's time to roll back. As for Google's "lower your target ROAS" recommendations, we don't treat them as optimization advice, since a lower target mainly lets Google spend more. We keep auto-apply off for them.

Step 6: Pace changes by conversion count

Let at least 50 conversions pass between target changes, ideally 60 to 90. For a small or mid-size campaign that works out to about once a month, and for a large one every week or two. Move the target 10-20% from its current value: 500% to 550%, or 1000% to 1100-1200% at most. A campaign with 10-15 conversions a month won't do much useful work on a ROAS target at all.

There are two ways to scale, and it pays to pick one. For more revenue, hold the target and raise the budget 10-20% every 50 or so conversions for as long as the campaign keeps spending it. For more profit, hold the budget and raise the target at the same pace until the campaign starts to slow down. Change the target and the budget on the same day and you'll never know which one did what.

Adjusting the target value within the range the campaign already reaches doesn't restart learning. Changing the strategy type does, for example adding a target to Maximize conversion value. After a change like that, leave the target alone for one or two conversion cycles.

Step 7: Handle promotions, seasons, and sudden shocks

Going into a sale with a target that's too high is the worst option, because the campaign can't catch the demand spike. On smaller accounts we don't switch a target on during a promotion. If one is already set, we lower or remove it and restore the old value afterward. Black Friday results never become the baseline for a target, since they're an outlier. Seasonality adjustments in Google Ads sound better than they work, because they hinge on a conversion rate shift nobody can predict well. Changing the target and budget by hand is more reliable.

After the season, a campaign with a target cuts its own spend as orders drop off. One without a target won't, so you have to trim the budget manually. When something big changes on the account, like downtime, a catalog swap, or a location turned off, raising the target won't help. Switch to a simpler strategy, such as Maximize conversion value with no target, and bring the target back after at least 30 new conversions.

Step 8: Log it, then check three times

Each change goes into a log with the date, the campaign, the old and new target, and the break-even number from Step 5. A month later, nobody remembers what they expected without it.

• Day 3: confirm the campaign is still serving and spending.
• Decision point: about 50 conversions after the change, which usually lands somewhere between day 14 and day 30.
• Compare cost, conversion value, and profit after ad spend over equal windows before and after, alongside campaigns nobody touched as your control group.

Leave out the day of the change and the most recent days. Conversion value keeps trickling in for days, so the latest week always looks worse than it will once it matures, and it can't tell you whether the change worked.

Step 9: The example from the video, a cut that didn't pay

On September 16, the store lowered the target on two product campaigns from 600% to 500%. The break-even math said spend needed to grow by about half. We compared the 32 days before the change with the 32 days after:

• spend went up 23%,
• conversion value went down 4%,
• ROAS fell from 613% to 480%,
• daily profit after ad spend fell 31%.

The remaining question was whether the change caused this or the whole store did. In the window after the change, Merchant Center rejected 860 products for a few hours over a landing page error. October on its own, after that outage cleared, still shows profit down 25%. Brand campaigns nobody touched brought in 65% more value per day over the same stretch, so demand was there. The extra spend bought traffic that didn't convert.

The call: the target goes back one step, to 550%, and a fresh measurement starts. If profit had grown while ROAS fell, the target would have stayed at 500%.

Rules at a glance

• Always compare target and result as percentages. Google stores the target as a multiplier, so 5.0 means 500%.
• Clean conversion value comes first: add-to-cart can't be a primary conversion, refunds get corrected, and consent-banner gaps get factored in.
• New PMax starts with no target. Add one at around 50 conversions, at the achieved ROAS or 10% below.
• A campaign that won't spend its budget has a target that's too high. Drop it below the real ROAS or remove it for a week or two.
• Beating the target while making money calls for more budget. A campaign spending its full budget below target needs a new target and the same budget.
• No ROAS target on brand, since brand protects existing sales. Different margins go in different campaigns.
• Before a cut, work out how much spend has to grow. Break-even ROAS = 100 ÷ margin.
• Change every 50 to 90 conversions, in 10-20% steps, one variable at a time.
• Lower or remove the target for a promotion and restore it afterward. Black Friday is never the baseline.
• Judge after about 50 conversions, over equal windows, by profit after ad spend, with untouched campaigns as the control.