Tally — Golf Scoring App

The Problem

Golf scoring apps are built for serious golfers.

Hole19, TheGrint, and their competitors assume you know your handicap, expect you to navigate course databases and GPS maps, and want you to pay a subscription. Heavy onboarding. Cluttered interfaces. Subscription before you've played a hole.

The Gap

These apps aren't built for casual golfers. If you play a few times a year, still use a paper scorecard, and have no interest in tracking your handicap, they're the wrong tool. Not too simple. Too much.

The Response

I'm that golfer. I couldn't find the app I wanted, so I built it.

Active scoring screen — 1 stroke in, last shot distance recorded

Active scoring: 1 stroke in, last shot distance recorded. Hit again or finish the hole.

The Design Process

Research

I didn't have user interviews. I worked with what I did have.

Competitive audit: I mapped Hole19, TheGrint, and TallyHo across onboarding complexity, feature depth, and Wear OS support, to see where casual golfers were being ignored and where the apps were actually breaking down.

Play Store review mining: The most useful signal came from competitor reviews. Wear OS sync failures appeared consistently: lost data, broken connections, abandoned features. This wasn't anecdotal; it was a pattern across real reviews.

Proxy user testing: Without access to a golf course during early ideation, I used golf video games to simulate round flow, working through hole transitions, score entry timing, and mid-round interruptions. Not perfect, but enough to work through the core interactions before writing any code.

What I didn't do: formal user interviews. I built on competitive evidence, real review data, and my own experience as exactly the person I was designing for. It's a real gap. I'm planning to close it in beta.

The Insight

Casual golfers don't need a better version of Hole19. They need a different product.

Wear OS was the opening. Every serious app has a watch companion, but real Play Store reviews confirm the execution is consistently broken. Sync failures, abandoned features, unreliable connections. For a casual golfer who wants to tap their wrist to add a stroke without pulling out their phone, no app is actually delivering this reliably.

That was the bet: something clean and simple that actually works on your wrist.

Design Decisions: What's In, What's Cut

The V1 filter was simple: would a first-time golfer understand this immediately? If not, V2.

In for V1

Stroke counting
Score vs par
Hole-by-hole summary
Round history
Watch companion (Wear OS)

Cut for V1

Handicap tracking
Course database and GPS maps
Tee box selection
Club tracking and swing analysis
Multiplayer

Every cut was to protect simplicity. Handicap tracking matters to a fraction of golfers. Course databases add onboarding friction for someone who just wants to start a round. V1 has to be simpler than a paper scorecard. That's the bar.

The Platform Decision

Tally is Android-native, and that's because of Wear OS.

Cross-platform frameworks like Expo don't support Wear OS. Going native was the only way to build what I actually wanted: a phone app and a watch app that work together. Both are built and working.

The phone app is currently in testing. The watch app is on hold while I wait for replacement hardware. The platform decision was right. The timing just got complicated.

Designing with AI: The Tastemaker Lesson

I used Gemini in Android Studio as a build tool, not a design tool.

Early on, before the Figma designs were done, I let Gemini generate the phone UI. The screens ran. But the UX was weak: generic patterns, default choices, no real design intent behind any of the decisions.

AI accelerates execution, but makes design decisions by default when there's no direction. Without a designer setting the standards, you get the average of what's been built before.

I stepped back. Designed the full phone UI in Figma first. Then used Gemini to implement against those designs. The result was better, and the gap between “AI-led” and “design-led, AI-implemented” was obvious when you held them side by side.

AI is fast at building. But it doesn't have taste. Someone still has to decide what good looks like. That job matters more now, not less.

The Work

Home: recent rounds list with date, holes, and score vs par. History without clutter.

Active round card pins to the bottom when a round is in progress: one tap to jump back in.

New round setup: pick 9 or 18 holes, name your course and round. No GPS, no database to search.

Hole setup with a circular par selector: par choices are few and predictable. The dial makes it fast.

Scoring screen at the start of a hole: two choices: hit your shot or finish. Nothing else in the way.

Between shots: the app tracks your walk and measures distance to your ball automatically.

After shot — stroke count and last shot distance

After the shot: stroke count updates, last shot distance logged. Keep going or finish the hole.

Hole summary with stroke-by-stroke yardage

Hole summary: result hero (“Par / Even, great work!”) with a stroke-by-stroke yardage breakdown.

Round summary — total score and hole breakdown

Round summary: total score, overall vs par, and a collapsible hole-by-hole breakdown.

Status

The phone app is largely complete and matching Figma designs. The watch app is paused pending replacement hardware.

Tally is in internal testing. Not yet public, but getting there.

All Work