Rize.gg MLBB Dashboard

When Rize.gg, a platform for finding gaming teammates, piloted its Philippines entry with weekly Mobile Legends: Bang Bang tournaments, the whole month came down to one number: would 150 to 300 people create accounts? I built the live dashboard the team watched it on, a single page that queries Google Analytics and PostHog directly, refreshes within minutes, and runs on no database at all. By day 28 the page showed 448 accounts, past the stretch goal with the championship week still to come, and nobody ever had to compile a report to know it.

Challenge

A pilot only teaches you something if you can tell whether it is working while it runs. This one had a hard target and a four-week clock, and the evidence lived in the wrong places.

• The truth was split across two tools. Philippines traffic sat in Google Analytics, which lags 24 to 48 hours; account signups sat in PostHog, which is near-real-time. The two counters never quite match, and neither tool is readable for someone who does not live in analytics.
• The audience was not technical. The person running the tournaments needed to check pace mid-event, often on a phone, and see "visitors" and "accounts created", not sessions, events, and HogQL.
• Pace mattered daily. With a 150-300 goal for month one, the team needed to know each morning whether the last tournament moved the number, not find out in a month-end report.
• A pilot cannot carry infrastructure. A short experiment does not justify a database, sync jobs, and a reporting pipeline that someone has to babysit after the month ends.

Approach

Build it for the person reading it

Everything on the page answers a question the pilot team actually asks. A goal bar shows live month-1 progress against the 150-300 target with the stretch goal marked. Four cards give the headline numbers in plain words, each labeled with how fresh it is. A week-by-week table mirrors how the pilot actually runs, registration days through tournament day, so each MLBB cycle reads as its own row with accounts, visitors, and peak day. The whole page wears rize.gg's identity, near-black background, mint accents, the Oxanium font, and holds up on a phone, because that is where it gets checked mid-tournament.

Two counters, shown honestly

Instead of pretending there is one true visitor number, the daily chart plots Google Analytics and PostHog side by side and says exactly why they differ: one runs a day or two behind, the other updates in minutes, and small gaps are normal. That honesty is what let a non-technical team trust the page instead of questioning it every time two numbers disagreed.

No database, no middle layer

The first version had a Supabase store and a nightly n8n sync. I deleted both. The Next.js server now queries the Google Analytics Data API, filtered to the Philippines through a service account, and PostHog's query API in HogQL, live on every page load. There is nothing to sync, nothing to go stale, and nothing to maintain when the pilot ends. Each source fails independently, so a broken key shows a small notice while the other series keeps rendering, and with no keys configured the page renders sample data so the UI could be reviewed before go-live. A basic-auth gate sits in front when credentials are set.

Results

Numbers as of August 4, 2026, day 28 of 31, with the championship week still ahead.

• The pilot beat its stretch goal in plain view. 448 accounts created against a goal of 150 to 300, with 933 new visitors from the Philippines and a peak of 243 in a single day, and the whole team watched it happen on the same page.
• The tournament rhythm became visible. Each weekly cycle shows as a spike that lands on tournament day: week 1 brought 165 accounts, the merged week 2-3 cycle added 283, so the team can see which events pull and plan the next one accordingly.
• Reporting effort went to zero. Nobody exports, merges, or formats anything. Opening the page is the report, and the signup numbers on it are minutes old.
• Nothing to babysit. No database, no sync jobs, no pipeline. Two API keys and a Vercel deploy, and if one source breaks the other still renders.

Tools Used

Next.js, React, Recharts, Tailwind CSS, Google Analytics 4 Data API, PostHog Query API (HogQL), Vercel