I built 8 free relationship quizzes in vanilla JS — no framework, no backend, no login

The technical approach makes sense to me. The bigger challenge is distribution, not architecture.

There are thousands of relationship quizzes online already, so I'd spend much more time figuring out why someone would share this one with a friend. If you can make the results feel insightful, accurate, or highly shareable, the lack of a backend becomes a strength rather than a limitation.

Curious, have you had any organic traffic yet, or are you still in the launch phase?

Shipping with zero backend complexity is underrated. The vanilla JS + client-side approach forces you to design for simplicity from day one. Most projects get bloated because backend feels "free" so people keep adding features. Here you have hard constraints. Did you find the no-login friction actually mattered (people skip quizzes because of signup), or did the relationship quiz topic itself drive engagement regardless? That'd tell you whether the frictionless design or the problem itself was the win.

No backend is the right call for this — the topics are personal and a login wall would kill the vibe. Single HTML files + share cards is a clean constraint. One idea: anonymous start/complete events only (never answers) so you can see which quizzes convert to the share card without storing sensitive data.

That's the question I genuinely can't answer yet.

From my analytics, most people take the quiz once and leave — which could mean the insights landed and they got what they needed, or it could mean the bar was so low to enter that there was nothing worth returning for.

My working hypothesis: friction isn't the enemy, but it has to be earned. Right now I don't have enough insight depth to justify asking someone to create an account. If I build something that genuinely tracks change over time — like "you scored X three months ago, here's what shifted" — then accounts make sense.

For now, treating each visit as potentially the only one forces me to make that single session worth something. Maybe that's a constraint that's actually useful.

No backend is the right default when the answers are this sensitive. You can still learn without storing responses: track anonymous start, complete, and share-card download per quiz, never the answers themselves. Add accounts only if people ask to compare results over time; otherwise you would trade the strongest trust feature for speculative retention.

The interesting challenge isn't building these quizzes without a backend—it's deciding whether removing friction creates more value than the personalization you'd gain from user accounts. I'd keep validating whether people return because the insights are genuinely useful or simply because the experience is effortless.