The "always take CPA on cold traffic" rule breaks at LTV $480

The "always take CPA on cold traffic" rule breaks at LTV $480

The advice: cold, untested traffic should always run CPA so the operator eats the risk. Mostly right — but the cutoff is mechanical, not a vibe.

The math, across ~40 tracked Tier-1 deals:
— CPA $180 flat vs RevShare 35% of NGR (net gaming revenue = stakes minus payouts minus bonus cost).
— RevShare overtakes CPA once player LTV (lifetime net value to the operator) clears ~$510 and breaks even around month 4.
— Below ~$480 LTV, CPA wins for the lifetime of the cohort.

The correction: "cold" is not the variable — retention is. If your source feeds sticky players (sportsbook regulars, low bonus-abuse), you're leaving 20-40% on the table taking CPA. The honest move is a hybrid: CPA $120 plus 15% RevShare, so you bank risk-cover and keep upside.

Run the first 200-300 FTDs on hybrid, measure month-3 LTV, then renegotiate to whichever pure model your actual cohort justifies.

Benchmark of the day: CPA beats RevShare only while cohort LTV stays under ~1.4x the CPA price.