Skip to content

All free lessons are open, no signup · e-books, courses and services are in preparation and not for sale yet

← Path: Selling online at a real profit

In depth · 7 min read

Are your ads worth it? Minimum ROAS from real margin

High ad sales are not profit — find your shop's break-even ROAS first, then judge each campaign

Many shops look only at "sales from ads" and feel happy, when after product cost, fees, shipping and ad spend they may actually be losing money.

What is ROAS

ROAS = sales from ads ÷ ad spend. Spend 1,000 baht and sell 3,000 baht → ROAS 3

This number shows sales, not profit, so you need to know the minimum ROAS your shop needs to avoid a loss

Find your minimum (break-even) ROAS

  1. Find your profit before ads as a % of the selling price = (price − product cost − fees − shipping you pay − packaging) ÷ price
  2. Minimum ROAS = 1 ÷ profit before ads (%)

Hypothetical example: sell at 500 baht with 300 baht of total costs → 200 baht left = 40% → minimum ROAS = 1 ÷ 0.40 = 2.5 A campaign with ROAS 2 is selling, but losing money on every order

Read the results honestly

  • Look over several days — at least 7; single days swing a lot
  • Repeat buyers: if new customers often come back, a first-order ROAS slightly below the minimum can be acceptable — but only with real repeat data, not hope
  • Sales the platform credits to ads may include people who would have bought anyway; compare total shop sales with ads on and off

Start with a small budget and set the rule in advance: "ROAS below the minimum for 7 days in a row = stop or change the creative" — then actually follow it

Key points

  • ROAS shows sales, not profit
  • Minimum ROAS = 1 ÷ profit before ads (%)
  • Set a stop rule before starting and judge over at least 7 days

Try with a free tool

Terms in this lesson

You finished this path 🎉

Questions or experience to share?

Ask in the community →

General knowledge only — not legal, tax or financial advice for your case. Check official sources or an expert before deciding.