Before Grab Was Grab, It Needed 40 Taxi Drivers to Give It a Chance
Grab’s early problem wasn’t getting people to download a ride app. It was getting enough taxi drivers to take it seriously. In Kuala Lumpur, the team behind MyTeksi started with 40 drivers and a standing invitation to sit down over coffee.

The app could match rides. It couldn’t make drivers trust it.
MyTeksi’s co-founders, Anthony Tan and Tan Hooi Ling, were working on a problem that had two sides: passengers wanted a more reliable way to find a ride, while drivers needed to believe a new dispatch system would improve their working day.
In a 2017 interview published by Grab, founding team member Aaron Gill said the company worked with a local taxi provider to recruit its initial 40 drivers. He recalled that many drivers didn’t yet have smartphones, so trying the app could mean buying unfamiliar equipment first.
The first users were also its toughest critics
Gill recalled that the team met the drivers every two weeks over coffee or teh tarik. Drivers weren’t shy about telling the team when things went wrong. Those conversations put a real weekly cost on what might otherwise have looked like a minor product problem: a tool that doesn’t work can cost someone a fare.
The important choice was to keep showing up. The drivers got a way to explain what wasn’t working, and the team got feedback from people actually trying to make a living with the app.
At Changi, onboarding happened at the taxi queue
Gill described how a version of that approach later reached Singapore. At Changi’s taxi queues, one team member would explain the app while another helped install it. It shrank the gap between hearing about the idea and actually trying it.
That is a different task from running a download campaign. The team was helping a driver start using a product, not asking them to remember an ad after the shift.
The lesson: choose a first group small enough to call
A founder building a service for merchants doesn’t need a polished nationwide launch to learn whether it works. Start with one cluster of operators who encounter the same problem. Sit beside them during setup, then measure which ones come back to perform the key action unprompted.
The number to watch isn’t sign-ups alone. It’s how many people keep using the service after the team stops helping them through every step.
