Experiments Over Opinions: Driving Growth at a Super App

5 MINS

For three years at airasia Super App my job came down to one discipline: replacing opinions with experiments. When you're driving growth through conversion-focused UX and personalisation, the most dangerous voice in the room is the loudest one. The second most dangerous is your own. Experimentation is how I keep both honest.

The funnel doesn't care what we believe

Everyone has a theory about why users drop off. Most theories are wrong, and the only way to find out cheaply is to test. I spent a lot of my time at airasia turning strong opinions into falsifiable bets, defining the roadmap around experiments that could actually move funnel performance, feature adoption, and revenue rather than just feel like progress.

The reframe that helped my teams most: we're not shipping features, we're running experiments that happen to ship as features. That small shift makes it normal to be wrong, which is exactly what you need to learn fast.

Personalisation is an experiment that never ends

Personalisation sounds like a feature you build once. It isn't. It's a standing experiment, what to surface, for whom, when, that has to be tuned against real behaviour continuously. The win isn't a clever model; it's a tight loop between what you show, how users respond, and what you change next.

I learned to be suspicious of personalisation that couldn't explain itself. If I couldn't articulate why a user was seeing something and how I'd know it worked, it wasn't a strategy, it was a guess dressed up as one.

Conversion-focused UX is mostly removing things

The biggest conversion gains I saw rarely came from adding something. They came from removing a step, a field, a moment of doubt. Conversion-focused UX is less about persuasion and more about subtraction, finding where the funnel asks too much and quietly giving that effort back to the user.

The discipline is resisting the urge to "add a nudge" when the real fix is one less screen. Every extra element is a tax on the conversion you're trying to earn.

How I decide what to test next

Not every idea deserves an experiment slot. I prioritise by where the funnel is bleeding hardest and where a result would actually change a decision. A test that can't change what we do next isn't worth running, however interesting it sounds.

Across discovery, definition, and distribution, my job was never to be right on the first guess. It was to build a machine that gets less wrong every week, and to trust the data over the meeting. Growth, it turns out, is just a long series of small, honest experiments that someone refused to skip.

From Incidents to AI Products: What Operations Taught Me About PM

Designing Search That Converts: Lessons From Travel

Chabi skipped presentations and built real AI products.

Chabi Anand was part of the April 2026 cohort at Curious PM, alongside 18 other talented participants.