You need ~300 settled conversions before EPC means anything

You need ~300 settled conversions before EPC means anything

Most affiliates make scale/kill decisions on 30-50 conversions, where the numbers are statistical noise dressed as signal — especially in betting, where whale concentration fattens the tails.

Why the threshold is high here:
— With the top 5% driving ~half of NGR, a single whale in a 40-conversion sample can double your apparent EPC, or their absence halves it.
— Clawbacks (3-30% depending on source) haven't settled at low volume, so early EPC is gross and optimistic.

Rough confidence ladder from tracked campaigns:
— Under 100 conversions: EPC swings ±50%+, decision-useless.
— 100-300: directional, ±25%, fine for kill decisions on clear losers.
— 300+ settled (post-clawback): stable enough to scale on, ±10-15%.

The practical rule: spend enough to reach ~300 settled conversions before trusting EPC to scale, but kill obvious losers earlier on cost-per-FTD, which stabilizes faster than EPC.

Benchmark of the day: trust EPC for scaling only past ~300 settled conversions — below 100 it's whale-and-clawback noise, not a number.