How do you validate a new offer before scaling?

[IndustrialAds]

Hey guys,
Quick question for those who are actively running paid traffic.
When you test a new CPA offer, what are your main validation criteria before scaling?

• Minimum spend before decision?
• EPC / ROI benchmarks?
• Number of conversions needed?

I’ve seen very different approaches and would love to hear what works for you in real campaigns.
Thanks in advance

 

[Rose agency]

I usually validate new offers using three simple steps:
A small test with sufficient data (typically 1–2x payout * 20–30 conversions).
Stable EPC, not fluctuating significantly between days.
Break-even ROI → positive after basic optimization (creative + flow).
Only if all three conditions are met should scaling be considered.
In reality, many failures don't lie with the offer itself, but with the quality of the account/BM and the stability of the advertising infrastructure. Using an account with a good history makes data analysis and scaling much more accurate.

 

[Hakeed49]

From my experience: at least 2–3× payout per spend, at least 10–15 conversions. I look not only at ROI, but also at the stability of EPC over the days. If it stays + and doesn’t “jump” — then we’re already scaling.

 

[John Blant]

I usually look at a combination of metrics rather than a single KPI.
First, I run a minimal test budget to see whether there are any conversions at all without strange traffic behavior.
Then I evaluate EPC/CPA in dynamics, not based on the first few days.
For scaling decisions, I need not just 1–2 conversions, but at least a small stable sample.
Sometimes an offer isn’t immediately ROI-positive, but CTR/CR can show potential, so I don’t cut it too early.

 

[John Blant]

I usually validate new offers using three simple steps:
A small test with sufficient data (typically 1–2x payout * 20–30 conversions).
Stable EPC, not fluctuating significantly between days.
Break-even ROI → positive after basic optimization (creative + flow).
Only if all three conditions are met should scaling be considered.
In reality, many failures don't lie with the offer itself, but with the quality of the account/BM and the stability of the advertising infrastructure. Using an account with a good history makes data analysis and scaling much more accurate.

Sounds logical, but 20–30 conversions as a “minimum test” is no longer really a quick test — it’s almost a mini scaling stage)