A live multiplayer product where everyone gets five HYPE, spends it in secret, and finds out together.
Open live SamePageAudienceB2C
Everyone gets exactly five HYPE and spends all five before locking in. Put all five behind one option, split 4 and 1, or spread them wider. Every point spent somewhere is a point you cannot spend somewhere else.
Choices stay private. When everyone commits, the room reveals the result together. There is no host waiting to press Reveal.
Five HYPE. Spend all of it. Your split is the opinion.
Reveal should release the answer, not leak it early.
Individual HYPE splits are never published. Each person can see their own allocation, but the room only reveals aggregate totals after everyone commits. In a two person room, SamePage warns before commitment that the totals can mathematically expose the other person's split.
Reveal stays neutral until the result is actually released: no early percentages, ranking, reordering, winner styling, or outcome story. Then the room counts, names the winner, releases the ranking, and settles the result. Genuine ties remain ties; an all tied result deliberately skips the celebration burst.
The final participant's commit resolves the room server side, so every client receives the same outcome instead of trusting whichever browser acted last.
SamePage originally asked everyone to react to every option with Absolutely, Down, Whatever, or Nope. It functioned, but it behaved too much like a survey: every option could stay alive, the scoring model contained hidden behavior around Nope, and the data structure could not preserve genuine co winners. The right move was not another visual polish pass. I deleted the interaction model.
A working feature was not the same as a good decision model.
Once SamePage became a real multiplayer product, small interface bugs became reliability problems.
Refreshing once erased unfinished HYPE. Reconnecting could leave a room waiting, and repeated Lock In could show an error after a successful commit. Draft persistence, a polling fallback and duplicate submission protection now preserve the decision.
A production smoke test found that Copy Link could inherit a deployment specific Vercel hostname. A fresh recipient reached a Vercel login screen even though the room itself was valid. Sharing moved to the canonical public domain and now refuses to silently generate a protected deployment URL.
Share now opens the browser or operating system's native share interface where supported, while Copy Link remains separate. If someone commits alone, SamePage says “You're in. Who else?” instead of pretending a result can happen with one participant.
Usability testing surfaced a different problem: first time users could reach the question flow before they understood what HYPE was. The explanation sat too low, the starter question treatment was not working well, and sharing needed a more native mobile path.
I rebuilt the landing for mobile. A short visual explanation introduces the five HYPE, private choices and shared reveal before the form.
The old starter questions became one lightweight “Stuck? Give me one” action that disappears once you start writing. Create was reduced to question, options, name, create. The product voice became shorter and more social without turning every line into a joke.
Clarity first. Personality where it helps.
Mobile first did not mean mobile only. Desktop became the responsive extension: aligned form columns, desktop only option seats, and one shared action row.
A unanimous 100% result overlapped its option label because the percentage column had only been sized for two digits. I corrected the layout for the actual maximum and checked it across screen widths.
The review also covered waiting, deleted rooms, keyboard use and Reduced Motion. Privacy boundaries and canonical sharing were checked separately from the interface.
Physical phone testing and testing with people who rely on screen readers remain unverified. No accessibility or security certification is claimed.
I replaced a working reaction model when it behaved like a survey, then rebuilt the introduction when people reached the form before understanding HYPE.
The live product keeps the tradeoff private until everyone commits, then lets the group find out together.
KAVRO organizes automotive commerce around completing the repair through one Job.
Prakhar Jain · Product Designer
3+ years in healthcare, enterprise software and digital products.
Open to relocation anywhere.
Explore selected work
I like sharing rough thinking, hearing useful disagreement and finding out what happens when an idea is built.
