How do I improve ROAS on Meta and Google?
| Meta | Google Ads | |
|---|---|---|
| Stores | 91 | 65 |
| Median ROAS on matched sales | 1.40 | 2.91 |
| Stores with ROAS under 1 on matched sales | 26% | 6% |
| Share of matched sales from first-time buyers (median) | 91% | 75% |
Stores with at least about $1,000 of spend on the channel in the 4 weeks and working tracking. Combined; no store is named.
Picture this
You run a skincare store. Meta says your ROAS is 2.2, so every Monday you raise the budget a little. Sales go up, the ad account looks great, and somehow there is less money in the bank at the end of the month.
Then you do the math on your margin: after product, shipping and fees you keep 40%. Your break-even ROAS is 2.5. And when you count only the orders you can match to those ads, ROAS is 1.9, not 2.2.
You were not scaling a winner. You were scaling a loss, a little faster every week. Nothing in Ads Manager would have told you, because Ads Manager does not know your margin and grades its own work.
The store is made up. The pattern is not: it is the most common thing I see when the bank balance and the dashboard disagree.
Google's median looks almost twice as good as Meta's in the table. Do not move your whole budget because of that. Google gets more returning buyers and more people searching your brand name, who were coming anyway. Meta finds people who did not know you. They do different jobs, so judge each against its own break-even, and against new customers.
1. Work out your break-even ROAS first
Every other step needs this number. Take your gross margin after product cost, shipping and payment fees. Break-even ROAS is 1 divided by that margin.
A ROAS of 2 is great at a 70% margin and a loss at a 40% margin. "What is a good ROAS?" has no answer until you know your margin.
2. Measure ROAS on sales that match real orders
Meta and Google each grade themselves. In our data Meta reported a median 12% more sales than the orders we could match to its ads, and Google 19% less. If you raise Meta's budget because Meta says ROAS is 1.6 while the matched number is 1.4, you can scale a channel that is under your break-even.
A spreadsheet works: export orders with their UTM source, add up each platform's orders, divide by its spend. We explain it step by step in what true ROAS is.
3. Turn off what loses money on that number
This is where most stores find money fastest. Sort your campaigns and ads by matched ROAS, and look at the ones under break-even after a fair amount of spend, for example 3 times your target cost per order.
- Pause the ones that have been under break-even for 2 to 4 weeks.
- Keep a small budget on new tests. A week is too short to judge most ads.
- Move the money to what is above break-even, in steps of about 20% so the platform does not restart its learning.
In our data, 26% of stores ran Meta at a ROAS under 1 on matched sales over the 4 weeks. By Meta's own numbers it was 21%. Some of that 5-point gap is spend that looked fine and was not.
4. Split new and returning buyers
Retargeting and branded search often show the best ROAS in the account, because they reach people who already know you. Some of them would buy anyway.
In our data, a median 91% of the sales matched to Meta came from first-time buyers, against 75% for Google Ads. Look at your own split. If a campaign's sales are mostly returning customers, test turning it down for two weeks and watch total store sales, not only the campaign's ROAS. Our guide on new vs returning customer ROAS shows how.
5. Send the real sales back to Meta and Google
Meta and Google bid on the sales they can see. When a purchase never reaches them (blocked by a consent banner, an ad blocker or a checkout redirect), the algorithm learns from fewer buyers and bids worse.
- Meta: set up the Conversions API, so purchases are sent from your server, not only the browser.
- Google: turn on enhanced conversions, or upload offline conversions for orders the tag missed.
Check afterwards that each order arrives once. A purchase sent twice teaches the algorithm the wrong thing.
6. Use target ROAS bidding, set from your real numbers
Both platforms can bid toward a ROAS for you. They aim at the value they can see, so set the target from your matched numbers, not their own.
- Google Ads, Target ROAS: Google predicts the value of each auction and bids to hit your average target. A target of 500% means 5 in sales for every 1 spent. Google wants at least 15 conversions in the last 30 days, and a higher target means fewer auctions and less total sales (Google Ads Help).
- Meta, ROAS goal and minimum ROAS: the ROAS goal aims at an average; minimum ROAS sets a floor. Both need purchase value optimization, and a target set too high can stop delivery (Meta Business Help Center).
Start a little below what the campaign does today, then raise the target in small steps. Raising it 30% overnight usually cuts volume more than it lifts ROAS.
7. Then improve the ads and the page
ROAS is also order value times conversion rate, divided by cost per click. You can move every part:
- Order value: bundles and free-shipping thresholds just above your average order.
- Conversion rate: the landing page should match the ad's promise. Send a product ad to that product, not the home page.
- Cost per click: test new creative every week. How to test so the result is real is in our ad testing guide.
When ROAS is the wrong goal
Higher ROAS often means less spend and fewer customers. A store at ROAS 4 with 50 orders a day may earn less than the same store at ROAS 2.8 with 120 orders a day. If you are trying to grow, watch profit after ad spend, and blended MER (all store sales ÷ all ad spend), next to ROAS.
FAQ
What is a good ROAS?
One above your break-even ROAS (1 ÷ gross margin). For reference, the median on matched sales in our data was 1.40 on Meta and 2.91 on Google Ads, but stores' margins differ too much for a single "good" number.
Why did my ROAS drop when I scaled?
The first part of a budget reaches the most likely buyers. Each new step reaches people a little less likely to buy, so ROAS falls as spend rises. Scale in steps of about 20% and watch whether total store sales rise with it.
Should I trust Meta's ROAS or Google's?
Inside each platform, use its own number to compare ads. Between platforms, use sales matched to real orders, because each one grades itself.
We make TrueROAS, which shows the ROAS of every ad on sales matched to real Shopify and WooCommerce orders and sends those sales back to Meta and Google. Steps 2, 3 and 5 are what it does for you.
Written by Rasmus Arvidsson, founder. Store numbers combined across stores, 5 Oct to 1 Nov 2026; no store is named.

