ScoreFloat
Designing a live product from zero — as the only designer
Type: Co-founded live product · Timeline: Ongoing · Stack: Live product · Cricket · Football · Tennis · Role: Co-founder, sole designer
THE PROBLEM
Watching live sport while working means constant tab-switching. Every time you check the score, you lose your place. Sports apps demand your full attention. Score websites are built for people who aren't doing anything else.
ScoreFloat is a different bet: a Picture-in-Picture score widget that floats above everything on your screen. You pin a match, close the tab, and the score follows you — in the corner, out of the way, always live.
MY ROLE
I co-founded ScoreFloat with a developer. Every design decision — the product experience, the scorecard system, the landing page, the research — is mine. Every build decision is his. That division is clean and it's made us faster.
WHAT THE RESEARCH TOLD US
Eight users. One finding that changed everything.
The core pinning mechanic scored 4.75 out of 5. Users understood it, liked it, came back for it. But one problem surfaced consistently: the score wasn't visibly updating in the PiP float. Users couldn't tell if what they were seeing was live or stale. They trusted the feature — they just didn't trust the data.
That single insight reframed the design problem. It wasn't about discoverability or onboarding. It was about trust at the moment that matters most: when a match is live and the score hasn't moved in two minutes.
THE DESIGN DECISIONS THAT FOLLOWED
The "Updated X seconds ago" timestamp The most direct response to the trust problem. Every live scorecard now shows exactly when it last updated. Not a spinner, not a pulsing dot — a precise timestamp. Users can see staleness before they feel it.
The three-zone card header State pill (Live / Upcoming / Results) + date and context + CTA. Each zone answers a different user question in order: what's happening, when, and what can I do. The left border colour carries the state signal pre-attentively — you know if a match is live before you read a word.
Separate CTAs per card state Live cards: Pin Score. Upcoming cards: Notify Me. Results cards: Pin Commentary — with a countdown showing how long commentary is available post-match. Each state has a different user intent. Giving them the same button regardless was a usability failure in the original design.
Six iterations. 55 to 84. The scorecard went through six design rounds tracked against a usability score. Starting at 55/100 on the original site. Landing at 84/100 by v6. The delta came almost entirely from three things: score typography dominance, timestamp visibility, and the state-specific CTA system.
WHAT IS TAUGHT ME
Designing as a co-founder is a different discipline. Every opinion needs evidence behind it because there's no brief, no PM, no stakeholder to arbitrate. You either bring data or you bring a well-reasoned argument. Usually both.
The trust problem also taught me something about where to look. Users weren't dropping off at onboarding. They weren't confused by the PiP mechanic. The problem was invisible — it only appeared when the product was doing exactly what it was supposed to do. You only find that in research.