I spent the past 2.5 years at Uber trying to educate our US-based product and engineering teams about Asia — at first while I led the growth of uberPOOL in China, and later when I led product strategy for the rest of APAC and went on to tackle the tough local nuances across 14 diverse markets. I was fortunate enough to have a front-row seat to Uber’s globalisation efforts, and to have fought alongside some of the most amazing operators who built a thriving transportation business from scratch.

In light of recent events in Southeast Asia, many people are wondering whether American tech companies can win in Asia — especially the group of Uber China alumni who later rejoined Uber APAC, only to be sold a second time. I believe that building product for a vastly different part of the world is always difficult, but not impossible — and that in today’s era of globalisation, it’s crucial that companies understand how to build for local nuances and handle the complexity of diverse customer bases.

While Uber’s original cultural values were criticised after the departure of Travis Kalanick (“TK” as he was affectionately known), I believe that they played a crucial part in the rapid hyper-growth that Uber saw around the world. Two of the ones that always stood out to me were “Celebrate Cities” and “Customer Obsession” — these two ideas were at the heart of how I approached product strategy for the region: think about how a local entrepreneur might build product.

Celebrate Cities

We’re passionate about cities. Everything we do is to make the city a better place. We see what the future of a city can be and celebrate the heroes and heroic efforts that make it a reality. Uber works best when we are woven into the fabric of the city with our mission to serve the people.

Ridesharing is an incredibly localised, nuanced business — it inherently deals with the complexities of urban development and transportation planning. The local cultural and geographic nuances in every city and country where Uber operated applied tremendous pressure on our product, whether it was figuring out how a motorbike driver in Jakarta might accept dispatches safely, or making uberPOOL more efficient in congested cities like Manila.

The latter was especially close to my heart — I was part of the initial launch of uberPOOL in Southeast Asia while I was still leading the China POOL team, and hosted a group of Southeast Asia ops in Chengdu to help them learn more about uberPOOL in May 2016.

At the time, the threshold for launching uberPOOL was that a city needed to be hitting 400,000 trips per week. Only two cities in Southeast Asia fit this criteria at first — Manila and Jakarta — but Singapore tagged along despite being much smaller. When I heard that Headquarters was not superpumped about launching uberPOOL in Singapore, I stepped up to fight for a launch there.

My belief that uberPOOL would work in Singapore, despite lower trip counts, was rooted in my experience with uberPOOL in China. Our primary metrics for measuring the success of uberPOOL in a city revolved around efficiency: how many uberPOOL riders actually got matched with other riders, and how much busier drivers got as a result of taking uberPOOL trips. The main contributor to efficiency was liquidity — the more trips being requested, the more likely it was that we would be able to match two riders together. Our ideal scenario was that two (or three) riders ordering an uberPOOL from the same office building after work, going to the same apartment complex, would share the entire ride home for a lower price but experience no inconvenient detours.

Far before the Southeast Asia operations team paid us a visit, we were struggling to identify why certain uberPOOL cities in China weren’t performing despite high trip counts. In particular, sleepy cities like Hangzhou or Suzhou were outperforming megacities like Guangzhou and Beijing, despite having similar or even lower trip counts.

I was stumped at first — until I compiled a heatmap of request and dropoffs in these cities. The trips in the cities looked something like this:

Visually, it became obvious why these smaller cities were doing better: liquidity was more than just a function of trip count, but a function of the city itself and how it was laid out. Hangzhou was not only compact, but everyone was going to the same part of the city; Guangzhou had multiple centers of economic activity and people were using uberPOOL to go everywhere in the larger city.

I took a look at maps for Jakarta and Singapore and came to the conclusion that despite Singapore having less than 400,000 trips per week, it would outperform Jakarta (which was doing far more trips). I took this hunch to some data scientists on the uberPOOL team, who ran a full simulation on Singapore — and found it would be one of the top performing cities in the world for uberPOOL.

When we launched uberPOOL in Singapore in July 2016, it was a resounding success. Riders loved the low prices, and drivers loved being busier — especially during the peak commuting hours where surge pricing made it so that drivers earned more. We even got a couple of employees later who were only interested in Uber after we launched POOL!

Customer Obsession

When determining what to do, we start with what is best for our customer. That can be hard, but we refuse to settle. Instead, we innovate our way out of the box. Many companies talk about being customer-focused, we’re customer obsessed. We are relentless in figuring out what matters to our customers and then doing everything in our power to deliver it.

The introduction of uberPOOL highlighted another problem in Asia for Uber: picking up riders was dramatically more complicated. The original Uber pickup experience was designed in the US, where most riders met their drivers curb-side — and they even were willing to cross the street if it meant a faster trip.

However, Asian cities looked quite different. For starters, most pickups didn’t happen curbside; rather, they happened in mega-malls or office buildings with complex layouts and multiple, designated pickup locations. Clearly, Uber’s original concept of “drop a pin on a map” wouldn’t work in here. Further, asking a rider to cross the street was a dangerous endeavour: major roads in Asian cities would require people to either cross multiple lanes of fast-moving traffic without a crosswalk or make a multiple-minute detour to find an underground passage or above-ground walkway.

In my presentation to new hires in APAC, I often joked about how our product and engineering teams knew more about the Caltrain station in San Francisco than any major pickup location in Asia. After all, senior execs at Uber would have to deal with getting picked up and dropped off at Caltrain everyday —  but they may have never visited Asia at all.

We tried to solve this problem in APAC by porting over a solution originally built for airport pickups — internally called “Venues.” Venues allowed operations teams to designate special geofences for pickup experience overrides, letting operators designate and label pickup locations. We eventually bastardised this solution to cover hotels and office bulidings with designated pickup points, shopping malls with basement and above-ground pickups, and apartment complexes encompassing multiple buildings and blocks where residents could get picked up. We let loose a team of operators to create these Venues across the APAC region, to the point where over 50% of the world’s Venues were created in APAC — despite us having a much smaller share of the world’s trips.

I later found it heart-warming that Grab added a similar feature into their app — letting riders in Singapore designate a pickup point in many shopping malls.

Can American Tech Companies Win in Asia?

To answer this question, I think we need to look beyond tech companies to the examples of Western companies who have won in Asia — Starbucks, Haagen Daz, and McDonald’s come to mind. I spent some early days of my career at PwC helping Western multinationals expand into Asia — and one of the frameworks we presented to our clients was the globalisation maturity framework.

The globalisation maturity framework highlights a key point for Western companies seeking to do business in Asia, or really, anywhere in the world: you must localise your products. Many tech companies are still stuck in the “export model” phase of their globalisation journey — where a headquarters-led team builds product to ship to the world — rather than living in the “originate model” phase where each individual region and market contributes actively to the global business.

Successful companies in China like Starbucks fully localised their operations somewhere between the regionalise and originate models; Starbucks in China was originally operated by a local Taiwanese company, who was able to quickly adapt to local market needs. However, Uber remained in the “export” phase of globalisation; product and engineering teams lived in San Francisco, and while a portion of the tech teams visited China, none of them actually lived in China long enough to truly internalise the problems that our users faced. Our operations and marketing teams in China were thus stuck with the products built in San Francisco, and had to hack around their limitations.

Going back to my approach to product strategy for a global company — “think about how a local entrepreneur might build product” — I firmly believe that American tech companies stand a chance to win in Asia. However, for them to be successful, they have to realise that there are tremendous differences in consumer culture and behaviour — and localise their product accordingly.