"Southeast Asia has scale. If you ever wanted to work at scale, you would go to Southeast Asia. We had 20 plus different services from transportation to food, shopping, medicine delivery, bill pay, movie tickets."
170M+ users across Southeast Asia
20+ integrated services in one app (ride, food, shopping, finance)
More rides per day than Lyft
More food deliveries than Uber Eats + GrubHub + DoorDash combined
The pressure: one UX for infinite jobs to be done
The Thesis
Physics First: Understanding Constraints Before Growth
Step 1: Define the physics — Market, product, model, channels (what currently exists)
Step 2: Find the biggest constraint — Where does the funnel break? Where is friction?
Step 3: Change one variable at a time — Don't explode the universe. Test one lever before moving to the next.
Step 4: Repeat — The world doesn't change dramatically. Stay rooted in reality.
Why this mattersMost founders try to change everything at once. Instead, understand what's actually working, identify the one biggest constraint, and optimize that piece.
60%
W1 retention target
4K
orders/day at start
The retention insight
60% week-one retention should flatten by week 2–3. If people keep coming back, the product works. This is the decacorn benchmark.
Playbook
Things That Don't Scale (But Validate Everything)
The WhatsApp driver test: Added 100 drivers to a WhatsApp group. They sold subscriptions in-car for cash. Manually verified conversions in backend. No code. Validated entire value prop.
The stadium hiring: Rented football fields. Boxes of phones & SIM cards. 60K drivers onboarded in weeks. Proved supply wasn't a constraint.
The credit card visual: Showed virtual wallet number on a picture of a credit card. People understood it instantly. Top-ups increased immediately.
The Typeform features: Tested new features (personality quizzes, forms) via Typeform. Deployed on web. No mobile release needed.
The GoPay driver incentive
Drivers as salespeople. When a customer had no digital wallet, the driver got an incentive to cross-sell. They explained the benefit directly in the car. Result: 60% of GoPay acquisition.
Why it works at early scale
Even 30 data points show the same trends as 30K points. Precision changes, not direction. So test with 30 people. Every idea is cheaper then.
Key Insight
One Lever Can Change Everything
"Using them as the salesman. You wouldn't believe how great of a salesperson someone can be when you were literally trapped in a car with them going somewhere. And so you have this captive audience, captive attention, you have someone who has the incentive to cross pay or cross sell someone into GoPay."
The driver became the distribution channel (not the app)
Real-world observation: drivers in jackets marketed Gojek to the city
Word of mouth was stronger than Facebook ads initially
One insight (use drivers as salespeople) = 60% of product acquisition
Rooted in physics: you already had drivers. You just hadn't used them.
Contrarian
What Most Founders Get Wrong About Growth
✗Big growth requires big product changesINSTEAD →✓ Use the levers you already have. The driver was always there. You just didn't use them to sell.
✗You need massive sample sizes to run experimentsINSTEAD →✓ Run experiments at 30 people. The trends are the same as 30K. Precision improves, not direction.
✗Automate everything as fast as possibleINSTEAD →✓ Do things that don't scale first. WhatsApp groups validate faster than code. Stadium hiring proves supply exists.
✗Copy is less important than product in growthINSTEAD →✓ Tie unfamiliar products to familiar concepts. A digital wallet shown as a credit card increases adoption immediately.