retention is the whole hinge, yeah. revshare only pays over time if the users you send actually stick around, so the choice turns entirely on whether your traffic produces sticky users or one-and-done ones. sticky traffic, revshare wins and keeps paying. churn-and-burn traffic, you are better off taking the pps. assess your real retention honestly before you pick, because revshare on churning traffic underperforms the pps you passed up, and pps on retaining traffic leaves the ongoing money on the table.
this is what the debate needs, the practical frame. in the abstract revshare-versus-pps just goes in circles forever, because there is no universal winner. for any actual operator it resolves fast, does my traffic retain, and can i wait for the revshare payoff or do i need cash now. answer those two about your own situation and you have your answer. from the management side the affiliates who choose well ask the practical questions rather than arguing the general case that never settles.
the revshare-versus-PPS question on cam and dating traffic comes down to it being a long-game-versus-short-game choice you can actually test rather than agonize over, because you can run a slice of traffic on each and let your real numbers tell you which pays more for your specific traffic. so test both on a slice rather than debating, let your numbers decide. so run a portion of your cam or dating traffic on revshare and a portion on PPS and let the actual results decide, because the models suit different traffic and testing settles which fits yours far better than argument, so the operators who choose well test both against their real traffic rather than reasoning about which should be better, which means the revshare-versus-PPS decision is testable, since splitting traffic between the models reveals which actually pays more for your specific audience and retention, and rather than agonizing over the long-game-versus-short-game tradeoff in the abstract you can measure it directly, making the practical test the resolution, because your own traffic’s behavior on each model is the only thing that definitively answers which one fits it, and a controlled test on a slice of traffic gives you that answer far more reliably than any amount of debating the general merits of retention-based versus immediate payout models.