Thank you for making this thread. I will keep this as constructive as possible, as I don’t take staff outreach for granted – you didn’t have to do this. Summed up, I have a suggestion for improvement, and believe change is necessary because while the current change satisfies are large percentage of users, it does not satisfy a large percentage of user types.
##Why this change needs improvement:
By purely looking at % of users, we’ve diminished the value of the statistic. One piece of valuable information we’ve neglected to pay attention to is the types of users this change affects. The largest type – the largest percentage – of players is the standard kind that plays games and little else. A number of other major types (dwarfed only because of the sheer size of ROBLOX’s playerbase) have been neglected.
While small in percentage, power users are a major driving factor for ROBLOX. To put this into perspective, the subset of these users that have front page games is even smaller, yet both of us would laugh at anyone who insisted that their numbers were so small that they didn’t matter. Power users are the current front page game developers, upcoming developers, and users who have enough brand loyalty that they’ve stuck around long enough to become power users – this group, while small in comparison to ROBLOX’s overwhelming playerbase, is responsible for attracting revenue from the largest group of players and providing a veteran community that can help teach new users.
This update is problematic for power users because this group of users has complex needs and may need to take advantage of more features to meet those needs. For instance, a power user may opt to play a building game with more people instead of a smaller server that’s closer to their region, as socialization is more important than ping for this genre. A power user may also decide to join a smaller server at the end of the list to ensure they join a new version of the game instead of an older, more-broken version that hasn’t phased out yet. Perhaps the most important of all is joining a game an acquaintance or friend of a friend is playing. I’m not friends with everyone in the Discord servers I’m in, but if they say “Anyone want to play X?”, I may decide to join in, and if they have their follow settings set to friends-only, I need to find them in the server list.
####Users in small, but dedicated, communities:
There are some niches that aren’t too large, but the users who do like that niche are amazingly passionate. As such, while the games they play may not amount to a large percentage of players, these games are still heavily utilized.
This update is problematic for small, dedicated communities because of the small amount of servers they have. For instance, if I play Survival 303 and land in a server with a toxic user who is ruining the experience for me, I’d want to join another server. If there are only two servers, and no one is my friend in the other server, I have to rely on matchmaking and a 50% chance of not landing in the same server again. This is even worse if one of my friends is in the toxic server, as ROBLOX will force me into that server to play with my friend.
####Users in groups:
I’m not sure how large the group community is, but I would imagine it would have to be pretty big. In addition to that, users in groups have very strong ties to their community and by extent to ROBLOX. To add onto that, half of the ads I see are for groups, so they have to be a large revenue source. This user type is probably almost as impactful as power users.
I believe the concerns for this group are known because of your comment on adding a permission, but this still does not alleviate all the needs of groups. Something @Polymorphic forgot was that it’s not only group members who use group places – in country, war, and miscellaneous roleplay groups, there is usually some form of allies/enemies (so much so that groups officially support these). These groups’ members may not be in the owning group and will not have permissions to view servers.
I’m unsure if this would make an impact on the PM botting problem even. If I were a botter, I’d have my legion of bots friend as many people as possible, so if I had a meager 1000 friends (which is very small in comparision to ROBLOX’s total playerbase) across these bot accounts, I could PM bot 1000 servers. The bots would likely target users on the forums, groups, and the users in their friends lists to still have plenty of people to PM bot and plenty of people to send friend requests to in order to grow their contact network like that “Shared with Google Drive” worm that started off small and then went on to affect people worldwide by recursively sending itself to its victims’ contacts.
##How this change could be improved:
I think there are better ways of approaching this issue. Individually patching the issues with per-game permissions and an option to join a game without a friend aren’t a good approach because you’re just duct-taping hole after hole at that point on a sinking ship. The feature’s core should be designed well enough that multiple use cases don’t need to be patched in.
As I posted here, ROBLOX could serve the list of servers without account PII. It would provide a rbxcdn headshot without username/userId, which couldn’t be traced back to the user, as well as the complete list of servers so the existing benefits of the server list are preserved. The only existing functionality we’d lose is mouseover names (though would still be able to identify acquaintances), but this could be remedied by allowing mouseovers on servers friends are in and highlighting their portaits with green so it’s obvious why mouseover works on some but not others, in addition to providing nice functionality. This only requires a single change to the core of the approach, instead of multiple reactive fixes. Even if there are some problems with this (e.g. would use up to many web requests to find users’ friends), this core idea should go to show that it’s reasonably possible to design a better solution.