Meta knew that defaulting teens into privacy settings would "[k]eep more people safe" at the expense of "risk to engagement metrics"

"1. Protect teens from unwanted DMs asap to keep them safe where bullying happens most 2. Protect teens' safety across additional interaction types while maintaining a simple privacy model" "Why not just do one? - We know from research that DMs is necessary for safety, though it's not expected. - We know from research that mentions and tags are expected. Matching user expectations keeps the experience simple. - We know bullying happens through mentions and tags and that private teens are more likely to be bullied this way. It's not a safe experience if we don't include mention and tags. - We'll have a frankenstein privacy model if we only include DMs, that will not match users expectations and therefore will not be simple." [6. Default all new teens into private + restricted interactions via registration, A. Default all new teens into private + restricted interactions via registration flow]: • Pros: "Keep more people safe from unwanted DMs, tags, and mentions" • Cons / Risks: "More risk to engagement metrics", "Could bloat graph", "People may not know a follow requester" • Mitigation: "Nudge people to follow someone they want to interact with", "Allow messages when requesting to follow someone" [6. Default all new teens into private + restricted interactions via registration, B. Upsell all teens via opt-in QP]: • Pros: "Less risk to engagement metrics" • Cons / Risks: "Teens may not understand the upsells as new users" • Mitigation: "Show after X number of days" [6. Default all new teens into private + restricted interactions via registration, C. Upsell all teens via opt-out QP]: • Pros: "Less risk to engagement metrics" • Cons / Risks: "Teens may not understand the upsells as new users PXFN may block" • Mitigation: "Show after X number of days"