Hello
I had an idea of creating an external system to track users across games, so to speak, by using publicly-available information about the user to make informed decisions about what special offers to show them.
Perhaps by inspecting their Roblox avatar, an algorithm could identify them as likely male or likely female to target them with a specific offer.
Or by keeping track of how a user interacts to specific types of offers (ie. how long did they spend looking at it before closing/buying, do they buy offers when the discount is above 50%, etc.), the system could share this data between a network of games to create a feedback loop.
All of the data it collects would either be public (like the playersā avatars) or collecting through in-game scripts. Is this allowed?
Iām sure this would be allowed if it was a single game, but given it shares this data between a network of games is casting a bit of doubt in my mind.
Thank you
While ROBLOX does demographical Ad Campaign targeting, I donāt think programming an algorithm to study players is ethical.
Issues:
Most likely a bannable offense.
Incorrect data from avatars (not to mention how hard it would be to analyze a catalog itemās target user).
You may not create a ROBLOX script for this, as scripts donāt have access to ROBLOX resources and APIs. (could be circumvented with proxy or external server you can host and use HttpService with)
Unless it was a custom in-game purchasing system, there isnāt really any API endpoints to get a Userās interest in a Catalog Item.
This would be an incredible amount of work to approximate gender from the combination of hats someone is wearing. Assuming it could be done, what is the benefit to the end-developer? If a game provides a different experience based off that data, is it not discrimination?
Permissibility aside, itās morally questionable to be tracking (mostly) kids around games to offer them deals that get them to waste more of their parentsā money - given we already have plenty of cash-grab high-contrast games that already try to exploit them as much as possible.
Your time would be better spent making your purchases more meaningful. Look into āBartle Player Typesā - that can be a useful pointer to determine what types of players exist and what factors they enjoy in a game. And just watch what performs the best in your game and optimize based off that
It would be difficult to tell which product a player might want the most. You would need to aggregate lots of data to first determine what āalikeā players buy most frequently. Then build a fingerprint for each new user and compare it to that. Maybe using a service like GameAnalytics could be helpful to accomplish this
Aside from the fact that a ranking algorithm (that serves content that Iād be interested in and that I donāt pay a monetary cost to access) is different from a system that OP suggested:
Weāre talking about young kids here that would be most gullible to this type of advertising, and donāt fully understand that theyāre being manipulated to buy the products theyāre recommended. I do see an issue with that.
If weāre talking about Googleās wider data collection practices used to recommend me ads on YouTube, then yes, I also believe thatās morally questionable.