May 2020. Alternative credit scoring for unbanked gamers in Southeast Asia — using gaming activity, in-app purchases, and in-game behaviour as proxy for creditworthiness, plugged into the Razer Pay ecosystem. Built a working prototype. Submitted to the Razer Fintech Digital Hackathon. Didn't advance past the first round.
the problem
70% of transactions in Southeast Asia are still done in cash. In a region where 27% of adults have formal bank accounts and 30% of unbanked adults globally are aged 15–24, the credit system excludes the people most likely to be gamers: young, cash-dependent, and digitally active but financially invisible.
438 million individuals in Asia have no credit history. Not because they're irresponsible — but because the data that would prove their creditworthiness has never been captured. Meanwhile, 66% of SEA SMEs can't access loans or lines of credit for the same reason.
Razer had just launched their bank and payment platform. They had the distribution, the user base, and the transaction data. The missing piece was a way to score the people already living inside their ecosystem.
what we built
PowerUp Credit — an alternative credit scoring tool integrated directly into Razer App, pulling gaming activity, in-app payment behaviour, and lifestyle signals to generate a credit score for users who have no traditional financial history.
Four things it unlocked:
Credit score — a transparent score generated from alternative data, exportable as an API for other financial institutions to use as a proxy.
Gamify + Level Up — the score improves based on in-game achievements and activity. Better behaviour unlocks higher credit limits and lower interest rates — mechanically aligned with how gamers already think.
Buy now, pay later — credits spendable across the Razer Pay merchant ecosystem.
Loans — short and long-term loans at rates that reflect the actual score: 0.5% to 12% APR, not the punitive flat rates the unbanked typically face.
the scoring model
The PowerUp score was built from four weighted components:
Alt Credit Score (65%) — the core signal, composed of: gaming data (35%), in-app payment behaviour (20%), utilities (15%), telco (15%), ride-hailing and online shopping (10%), word cloud analysis (5%).
Activity (15%) — payment timeliness tracked at 30, 60, and 90 day intervals; negative balance history.
Frequency (10%) — in-game credit top-up patterns, defaults, late payments.
Postal Code (5%) — a proxy for alternative collateral, using location as a signal for assets owned by the user or their family.
The interest rate formula was explicit and user-facing:
adjustedInterestRate = (1 - metascore) × (max_interest - min_interest) + min_interest
A metascore of 0.7 → (1 - 0.7) × (5 - 0.5)% + 0.5% = 1.85% on a short-term loan. The better you play, the less you pay.
architecture
Built on AWS Elastic Beanstalk, pulling from game publisher APIs (Dota 2, League of Legends, and others). The bank layer ran on Mambu (banking-as-a-service). Scoring signals included: HowOftenGamesPlayed, ProbabilityOfHavingJob, HasInternetLag?, DoesUserHaveAComp?, RiskProfileInGame.
We shipped a working prototype with a live test account for the submission.
business model
Four revenue streams:
- Transaction fee — 2.5% on Razer Pay transactions
- Credit Score API — $3/report for unbanked users; $1.5/record for real-time updates to complement existing scores
- Interest on loans — 0.5% to 12% APR
- Data subscription — previously uncaptured datasets on the unbanked, sold to financial institutions
competitive positioning
Versus Credolab and Lenddo, PowerUp was the only player with all four: real-time events, transparent metrics, actionable behaviour change, and adjustable loan interest rates. The gamification mechanic — where improving your gameplay literally improves your credit score — was unique.
market
- 438M unbanked individuals in Asia
- 1.5B total gamers in Asia Pacific
- $80B untapped credit gap in Southeast Asia
- SEA gaming growing at 30% Y-o-Y; average 64 minutes of daily game time in APAC
the team
Built with Daryl Lim (Innovation Consultant at Padang & Co, exited AI founder), Shan Lin (Innovation Manager at Dun & Bradstreet), and Wen Kai Chia (IT Consultant at MOM, full-stack developer). I led product and business.
what happened
125 teams entered. Top 20 advanced to the semis. We didn't make it.
The first round was screened by a marketing representative — not a fintech person, not a credit scoring person, not someone equipped to evaluate a scoring model, a Mambu integration, or the logic of using gaming behaviour as creditworthiness proxy. The idea was rigorous. The judge wasn't.