Launching a Bilingual Website for a Golf Coaching Business
7 inquiries and 2 enrollments in the first month
A golf coach in Vancouver needed new students from two audiences, the Korean community and English-speaking golfers. As his marketing lead, I planned, built, and now measure a bilingual website that is the only path to his KakaoTalk and WhatsApp inquiry channels.
Date
TYPE
my roles
Marketing Lead
Web Build
Analytics
The Problem and Key Decisions
Problem
A golf coach in Vancouver needed new students from two audiences, the Korean community and English-speaking golfers. He had no website and no way to tell which marketing brought people in.
Key decisions
Meet each audience where it already talks. Korean customers reach the coach through a KakaoTalk open chat, and international customers reach him through WhatsApp. The site is the only path to both, so every inquiry can be traced back to it.
Build for search, not only for social. I created bilingual local SEO pages, including regional and junior lesson pages, and structured the content so AI search tools can quote it (GEO). Search keeps working after a post stops getting views.
Verify the measurement before trusting it. A weekly agent report showed zero contact clicks while real inquiries were arriving. I treated inquiries as the source of truth until I fixed the tracking.
What I did
Built the site with Claude Code and deployed it on Vercel in Korean and English.
Set up PostHog and Vercel Analytics, plus a scheduled Claude Code agent that emails a weekly report with prioritized fixes.
Acted on the reports. I fixed the tracking, lowered a section-visibility threshold that was set too high, added UTM tags to social links, rewrote copy and buttons, and strengthened the SEO pages.
Results and What’s Next
Results
In the first month, the site generated 7 inquiries and 2 enrollments, a family of three and one individual.
Google search kept bringing steady traffic after an early social spike faded.
The weekly agent report surfaced a tracking gap and missing UTM tags, which I fixed.
What’s next
The sample is still small, so I treat these numbers as a baseline rather than proof of what caused them. With UTM tags now in place, my next step is to compare inquiry rates by channel and by page before I test any paid ads.
Skills this shows
Landing pages · SEO and GEO · Analytics and event tracking · Funnel analysis · Bilingual marketing
Tools
Claude Code · Vercel · PostHog · KakaoTalk · WhatsApp