When to scale your ad budget: three tests before you add spend
Margin turns positive, and every founder wants to go all in. Vegas loves that founder. I try not to be one.
Every UA budget I build runs like a casino floor, not one table. And no table earns more chips just because it looked hot for a night. A channel, campaign or geo gets more budget only when it passes the same three tests, every time.
Test 1: Enough hands played
Under 30 to 50 conversions, you do not have a result. You have noise.
Confidence in a number does not grow with your sample size. It grows with the square root of it. To double your certainty you need four times the data, not two. That is why a great first weekend proves almost nothing.
Test 2: Clearing the real number, not a nice one
Your breakeven ROAS is 1 divided by your margin.
| Margin | Breakeven ROAS |
|---|---|
| 50% | 2.0x |
| 60% | 1.67x |
| 75% | 1.33x |
At a 60% margin, 1.67x is the point where you stop losing money, not where you start winning. Anything below it is not performance. It is a slower bleed.
Use the ROAS for your real payback window. A day-7 number that clears breakeven means something very different from a day-180 number that clears it.
Test 3: Held twice, not once
One good week is a lucky night. Two in a row is a table that is actually paying.
Seasonality, a featured placement or one viral creative can lift a single week. Two consecutive periods above breakeven, with enough conversions in each, is much harder to fake.
Fail one, and the table stays dark
No exceptions, and no "it feels like it is about to pop off."
And winning at one table does not buy a seat at the next. Every channel, geo and campaign runs this test on its own. A strong Meta account says nothing about TikTok.
Where founders go wrong
Most founders skip straight to the bet. Revenue is up, so they scale the whole account. That is betting on the mood instead of running the math.
The other mistake is the reverse: a ROAS far above breakeven that nobody scales, because it looks too good to touch. If marginal ROAS is well above breakeven and retention is stable, you are probably underspending.
The move
- Write your breakeven line on the media plan: 1 divided by your contribution margin.
- Set a minimum of 30 to 50 conversions before any channel's result counts.
- Require two consecutive periods above breakeven before adding budget.
- Scale in steps, for example 25% to 30% at a time, and re-run the three tests after each step.
- Run the tests per channel, per geo. Never let one winner pull the whole account up.
Bottom line
Three conditions. Not a feeling. That is the whole formula.
FAQ
How do I calculate breakeven ROAS?
Divide 1 by your contribution margin. A 60% margin means a breakeven ROAS of 1.67x. Below that, each sale from ads loses money.
How many conversions do I need before I scale a campaign?
A practical minimum is 30 to 50 conversions. Below that, the result is mostly noise. Because confidence grows with the square root of the sample, doubling your certainty takes four times as much data.
How fast should I increase ad spend?
In steps. Raising budget by about 25% to 30% at a time, and checking that the channel still clears breakeven after each step, avoids breaking what was working.
Should I scale all channels when one is profitable?
No. Each channel, geo and campaign should pass the tests on its own. Success in one channel says little about another.