Context beats fragmentation. Brand is easy to measure, hard to break through. And AI should serve the human, not the model.

Wong started as a psychology major trying to understand people. He worked at Holland Partners, a consultancy born out of the UK, on brands and communications, connecting with Yahoo and eBay in the early 2000s. That led to about a decade at Yahoo in the search business on the marketing and research side, then Pinterest to build a business analytics practice as it started connecting with advertisers, and years at BuzzFeed and Vox Media studying how content and podcasts flow with consumers. Uber let him do something new.

There are a lot of metrics out there. What drew him to Uber is the ability to close the loop for advertisers. With an opt-in user base and first-party data, Uber can see the sale itself, so it can look at incrementality for advertisers specifically, while still partnering with third parties so marketers can grade its homework.

Growth comes back to context and changing media behaviour. The first 50 of the last 70 years were about leaning back, and advertisers capitalised on what we watched. Mobile made us lean in through the device and the feed. The last 10 years, gaming, the metaverse and augmented reality, made it about being inside it. Now we're in a fourth wave where the digital experience has to create consumer control, and a utility like Uber is instant enough to impact real life. It is hard because the way consumers interact with media changes so often.

Uber serves about 200 million consumers monthly, over 15 billion trips in 2025, and about 45 million loyalty members, which builds deep everyday context. Wong measures four things. On attention, Adelaide shows Uber is about 40 percent more attentive than typical digital inventory. On expansion, coupling Uber with linear TV gives a 56 percent increase in incremental reach. On connection, brand lifts run about 10 points higher on favourability. On transaction, coupling offers with sponsored listings gives an 8X greater return in attracting new users than either alone.

Around 50 percent of open-web traffic may be real or fake. Because you have to be a real person to use Uber, and you have to be delivered a ride or an order, it builds trust with the consumer and means an advertiser need not worry whether that traffic is quality. As a connective commerce platform, Uber is doing something very different as a media platform.

Performance will always matter, but brand love and what a platform can help a brand do at the upper funnel will remain critical, and that is the new way marketers want to work with Uber.

Wong pushes back on the idea that brand is hard to measure. Working with Adelaide, Uber built a custom ad unit. Its ads perform about 40 percent better on attention, and hundreds of Kantar studies showed brand consideration lifts around 10 percent, roughly three times the global average. By correlating attention with brand scores, chief marketing officers can see the consumer both implicitly and explicitly, a new way to talk about brand.

An ad is viewed for about 100 seconds during a ride. With Pizza Hut, introducing a hand-crusted sourdough pizza, strong creative held attention for three minutes and, paired with sponsored listings and the pizza context, drove a 26 percent increase in new-to-brand users. On a 20-minute ride, engaging content plus the right context nudges the rider to decide tonight is the night for pizza.

Brands want to be around events, and nothing says events like source destination, a concert, a game, a date. Treated as a signal, a chief marketing officer can connect with a fan heading to a place without paying a content premium. Destination offers, launched at Cannes, are what Wong thinks is next, and the aim is to make the offer an elegant, native part of the experience.

The same AI and machine learning that deliver a ride with logistical precision power what ads to serve and how to connect with the consumer. Wong's challenge is capturing that magic in the algorithms and expressing it to marketers so they can see the effectiveness. He admits it is a question he is still working on.

What is different for Uber is that its algorithm is not driving the feed, it is delivering on a need. The service works by what it delivers for real life rather than by driving more engagement. That source signal lets marketers see a different kind of context than they might from algorithms on other platforms.

Research has changed because consumers change so fast, but what endures is understanding why a consumer connects with a platform and with media. Podcasts rose because people wanted a break from the endless feed and depth from trusted voices. In the wave of AI, consumers are looking for more human connection and pushing back for a reason, like the Luddites, textile artisans defending their connection to the trade. The pandemic accelerated the switch to audio and the search for a tribe.

Governance becomes a thing when there is a lot of data, deciding what to use and when. Wong uses AI and chatbots to find the right consistent queries, and stresses hygiene for LLM models. He argues a smaller, constantly refreshed data set often beats crawling the whole web, which can be outdated and can hallucinate.

New tools make everyone a researcher, everyone an expert, everyone a coder, which worries Wong about the loss of expertise and the need for curation. A great baseline from an LLM is useful, but knowing the right question to ask, being a great listener and having a real base of knowledge is what lets an expert build on it. AI cannot yet replace a subject matter expert, it is an assistant.

For anyone entering marketing, media or tech, Wong offers two things: be curious, because the space changes so fast, and have really cool friends. His best friends are always doing cooler things than he is, and studying how they connect with media, each other and culture makes him a smarter researcher, living life vicariously through them.