帖子

Ads Manager said ROAS 2.6. That week the ads lost money.

Break-even ROAS = 1 ÷ your margin after product, shipping, payment fees and refunds. Then check total revenue against a no-ads week.

I wrote out the full math, test budgets included.

Link in the first reply.

第一条评论

https://autowhisper.xyz/en/playbook/ad-math?utm_source=x&utm_campaign=playbook

X 长文版

Ads Manager Said ROAS 2.6 and the Ads Still Lost Money That Week

A lot of Shopify owners I talk to have the same problem. Ads Manager shows a ROAS that looks fine, they think about doubling the budget, and at the end of the month the bank balance has barely moved.

Usually there's nothing wrong with Meta. Nobody did the arithmetic. Here's the arithmetic, using one made-up store. Every number below is hypothetical, not a real customer. Swap in yours.

The store: dog joint supplement, $40 a tub, free shipping.

1. What one order really earns

• Price: $40.00
• Product + packaging: −$9.00
• Fulfillment + postage: −$8.00
• Shopify Payments (Basic, 2.9% + 30¢): −$1.46
• Refunds (assume 5% of orders): −$2.00
• Profit per order before ads: $19.54, a 48.9% margin

Easy to miss: international cards cost an extra 1% on Shopify Payments, and if you use another payment provider Shopify adds 2% on Basic (https://www.shopify.com/pricing).

2. Break-even ROAS

Break-even ROAS = 1 ÷ margin = 1 ÷ 0.489 = 2.05.

At a 30% margin you'd need 3.3. Same ROAS, completely different business.

3. The most you can pay for a customer

Break-even cost per purchase = $19.54. If you want $6 profit on the first order, your target is $13.54, which is a target ROAS of 2.95.

Why ROAS 2.6 lost money

Meta counts a purchase within 7 days of a click or 1 day of a view. Some of those people were going to buy anyway, from your email or by searching your name, and Klaviyo and Google claim the same order.

So look at the total. One hypothetical week:

• Meta spend $1,400, Meta-reported revenue $3,640 (ROAS 2.6)
• Total Shopify revenue $4,200 (MER 3.0)
• In weeks with ads off, the store does about $1,500

Extra revenue from ads: $2,700. Real return: 2,700 ÷ 1,400 = 1.93, below break-even.

In cash: $652 left after ads, versus about $733 in a week with no ads. Ads Manager said 2.6 the whole time.

First order loses money? Do the repeat math honestly

If the average customer places 1.6 orders a year, 12-month value is about $31. But use your measured repeat rate (wait for 90 days of cohorts), and remember you pay for the customer today and get paid back over months. When cash is tight, bid on 90-day value.

How much a test costs

Test budget = events you need × cost per event.

Assume $0.70 per click and 3.5% conversion, so about $20 per purchase.

• ~150 clicks (~$105) to judge whether the hook works
• ~10 purchases (~$200) to know the rough cost range. With 3 purchases, the true cost could be anywhere from $7 to $97.
• Stop a creative that's spent 3× break-even CPA with zero sales. If it really were at break-even, there's only about a 5% chance it would show nothing by then.

And count what the creative itself cost to make. That's part of the test.

If you want a fixed number for that line: in AutoWhisper a full video is 52 credits. The Meta ads it builds start paused, and you decide when they run and how much they spend.

Full write-up with tables: https://autowhisper.xyz/en/playbook/ad-math