I normally start with registration to build data, then move toward deposits when there are enough conversions for the algorithm to optimize effectively.
i’d optimize toward the deepest event you can get enough volume for, usually first deposit or qualified deposit. registrations are useful for testing, but they can be misleading if the users don’t retain or generate revenue.
deepest event you can get enough reliable data on. optimizing for registrations because they're cheap can train you toward traffic that's very good at registering and terrible at producing revenue.
but going straight to a deep event with almost no volume can leave the system with nothing to learn from. i'd move deeper once there's enough conversion density to make that signal usable.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.