5 years back when card tokenisation was introduced every "fintech" had a problem - no-one had the scale to convince a bank to allow tokenisation in their apps. Even Google had a chicken and egg problem. A thread https://twitter.com/MelissaFrakman/status/1315837359909416962

Tokenisation had two major problems to roll out
1. No massive app platform to standardize and introduce it on
2. Acceptance - NFC / custom ecom checkouts (masterpass, V checkout) adoption was poor

Enter UPI (4 years back)
- Fintechs could leverage banks authentication and liabilities in a unique 4 party model. Large app platforms emerged - GPay, PhonePe
- explosion of QR acceptance
- people comfortable paying with their phones

Meanwhile, V/MC did the ground work
- encouraged banks to build card tokenisation
- BharatQR and UPI QR merged specs to accept card payments
- no extra factor for contactless txns less than 2K (under NFC)

Meanwhile, Google built checkout by GPay and PhonePe built checkout by PhonePe for UPI. Everyone kept wondering why anyone needs this when it's just UPI.

And here we are 2020, with GPay introducing tokenisation with 3 banks (SBI credit cards, axis and kotak debit and credit) and more will follow - all as part of the "Checkout by GPay"

As things stand today, card tokenisation is a much better CX than UPI (atleast for txns below 2K). Supports debit and credit cards. Large app platforms are available to roll out and acceptance is sorted more or less (bharatQR, NFC POS, custom ecom checkouts)

UPI's regulatory arbitrage is all but gone. V/MC have played a great waiting game and slowly built key blocks. Next move - NPCI. Will UPI add credit? Fingerprint auth? Will they improve CX, success rates, chargeback management?

This is real twist in the story of UPI winning over V/MC. Time will tell but you know who's good at HA tech, uptime, bank incentives, chargebacks, frauds. <End>