Building Data-Driven Marketplaces and Automated Logistics with Natarajan Subbiah of Uber Freight

Key Takeaways:

‍

(02:30) Marketplaces thrive on data to balance supply and demand efficiently.

(03:30) Uber Freight”s introduction of upfront pricing to simplify logistics.

(07:20) Liquidity is the key metric for marketplace success.

(07:35) The easiest way to grow a marketplace is to constrain it on multiple dimensions. 

(12:10) 80 to 90% of complete goods or supply parts for goods get moved on trucks.

(17:22) What scaling automatization means for the freight industry.

(22:14) Over time, about 80% of freight will be moved in a highly automated way. 

(26:59) Centralized vs de-centralized teams, pros and cons.

‍

[00:00:00] Natarajan: I don't think there's a clear winner between centralized and decentralized teams, at least in my experience. It's not obvious that centralized teams are inherently more efficient or effective, nor that decentralized teams are lacking. It often comes down to how quickly teams can gather context.

‍

[00:00:42] Anthony: Welcome to Data Masters, the podcast where we delve into the minds of data-driven leaders shaping the future. Today, we're pleased to host Natarajan Subbiah, Chief Product Officer at Uber Freight. Natarajan is a seasoned product leader and marketplace expert with a proven track record in scaling businesses and driving innovation. Under his leadership, Uber Freight has become a major player in logistics, managing billions of dollars in freight annually. In this episode, we'll explore his insights on data science, marketplaces, and the future of logistics. Welcome to Data Masters, Natarajan.

‍

[00:01:31] Natarajan: Thanks, Anthony. Happy to be here.

‍

[00:01:33] Anthony: Great! Let’s start with the concept of being data-driven. Uber is known for being a highly data-driven organization, and I suspect that's core to your role. However, not every decision can be made entirely with data; intuition and experience also play a role. How do you balance these approaches in your work?

‍

[00:02:20] Natarajan: That’s a great question. In Silicon Valley, we’re accustomed to relying heavily on data for decision-making, and that’s been my approach for the past 20 years. Marketplaces, by their nature, are data-heavy businesses. At Uber Freight, data is central to decision-making, whether it’s pricing, supply and demand, or logistics optimization.

‍

However, there’s a fine line between being data-informed and overly data-dependent. Sometimes teams use data as a crutch, waiting for perfect information instead of relying on intuition and risk assessment. For example, if the risk of an A/B test is low, there’s no need to delay a decision for months to achieve statistical significance. It’s about striking a balance and learning when to act quickly.

‍

[00:05:12] Anthony: That makes sense. Data-driven decisions can also help remove subjectivity and bias, especially in hierarchical organizations. But as you said, it’s not a one-size-fits-all solution. Let’s pivot to marketplaces. You’ve built and scaled several during your career, including Uber Freight. What are some key challenges in building marketplaces, and how does data science help address them?

‍

[00:06:50] Natarajan: Marketplaces are challenging to build but highly defensible once established. The key metric for any marketplace is liquidity—how well supply and demand align. Data helps identify where to constrain or focus efforts, like targeting a specific city or market segment.

‍

For example, if you’re launching a dating app, you might start in San Francisco, ensuring supply and demand match in that city before expanding. Similarly, we use data at Uber Freight to optimize specific freight lanes, making it easier to build liquidity in neighboring lanes. Identifying tipping points and prioritizing expansion areas are heavily data-driven decisions.

‍

[00:11:13] Anthony: Many listeners might not be familiar with Uber Freight. Could you explain what sets it apart as a marketplace and its role in logistics?

‍

[00:11:41] Natarajan: Uber Freight is the largest 4PL in the U.S., managing about $20 billion in freight annually. Essentially, we help companies move goods more efficiently by providing both software solutions and marketplace services. For example, we might manage a retailer’s $3 billion freight spend, offering tools and expertise to reduce costs.

‍

Logistics has traditionally been a human-driven industry. At Uber Freight, we’ve pioneered innovations like upfront pricing and real-time load tracking. By leveraging data, we optimize networks, pricing, and supply chains, moving the industry toward greater automation and efficiency.

‍

[00:22:13] Anthony: Looking ahead, what trends or disruptions do you foresee in logistics? And how might technologies like large language models play a role?

‍

[00:22:43] Natarajan: Freight is a massive, complex industry. The next wave of innovation relies on digitizing operations, enabling automation and resilience. While LLMs are promising for tasks like invoice processing and SOP adherence, the real game-changer will be integrating these tools with robust data platforms. Over time, we expect 80% of freight operations to be highly automated, with humans focusing on exceptions.

‍

[00:24:02] Anthony: Shifting gears, how do you manage high-performing decentralized teams? Any strategies for balancing collaboration and efficiency?

‍

[00:24:48] Natarajan: Context is key, especially in a complex industry like logistics. For new team members, I encourage shadowing operations teams or directly engaging with customers to understand pain points. We also emphasize co-location for teams working on specific projects to reduce friction from time zones. Ultimately, whether centralized or decentralized, success depends on maintaining strong communication and context-sharing.

‍

[00:29:18] Anthony: Lastly, how do you stay updated on advances in data science and technology?

‍

[00:29:34] Natarajan: Personally, I’m an early riser and dedicate time each morning to reading industry updates and research. At Uber Freight, we also host learning lunches where team members share insights on various topics. These sessions foster curiosity and knowledge-sharing, helping us stay at the forefront of innovation.

‍

[00:31:38] Anthony: Thanks so much for joining us and sharing your insights, Natarajan.

‍

[00:31:44] Natarajan: My pleasure. Thanks for having me, Anthony.

‍

‍