Offers get picked by CPA rate, and that is the first mistake. The rate says what you are promised; what you receive is decided by the conditions underneath it. Five things that determine the outcome, and the rate is not first among them.
1. Minimum deposit
CPA pays for a deposit above a threshold, not for a registration. The gap between a $5 and a $20 qualifying deposit is a difference in conversion of several times over — someone willing to try a small amount and someone willing to put down twenty are different people.
A high rate behind a high threshold is almost always worse than a middling rate behind a low one. Count the rate multiplied by the share who get there.
2. Hold
The period the programme keeps your money before paying out — a week to a month. During the hold you pay for accounts, proxies and software out of your own pocket, which makes hold a working-capital requirement rather than an inconvenience.
3. Geo
Rates differ several times over between countries, and that is not a gift: an expensive geo means expensive traffic and strict quality requirements. Look at the pairing of «rate / where I would actually find this audience», not at the number.
The easiest place to find an audience for a given geo is topical chats — language and timezone line up by themselves there.
4. Shaving
A programme can fail to «see» some of your deposits. You cannot verify this directly, but the signs are there: conversion dropping while traffic stays the same, your numbers diverging from their report, payouts delayed without explanation.
The only defence is keeping your own statistics from day one. Without them you cannot tell shaving from bad traffic, and you will fix the wrong thing.
5. The product
The least-checked item and the one that matters. Register yourself and walk the path your referral will walk: how many steps to a deposit, how many payment methods, what happens on withdrawal. A product where people get stuck on step two will eat your conversion no matter how good the traffic is.
At 1win Partners revshare starts at 50% and CPA goes up to $250 per player. The first is a starting percentage, the second a ceiling reached by people whose traffic clears the quality and geo bar. Ask not «what is your maximum» but «what will it be for my geo at my volume».
Testing an offer without burning budget
Take a small volume and one geo. The goal of the first run is not profit but your own numbers: contacts per registration, registrations per deposit. With those two figures you can compare offers; without them you are comparing other people's promises.
How to work out acquisition cost step by step is in cost per lead from Telegram.