How Whatnot Built a Global Marketplace
Grant LaFontaine, co-founder of Whatnot, discusses how the platform transformed live commerce from a niche Asian trend into a major marketplace by focusing on user experience rather than market mechanics. The company has built a $16+ billion annualized GMV business empowering small sellers across 100+ categories while maintaining human-centric connections between buyers and sellers.
Summary
The episode opens with Grant's origin story: selling a holographic Pokémon card on eBay in third grade for $10, which sparked a lifelong interest in online commerce. After working at YouTube and Facebook, he and co-founder Logan reconnected in 2019 with a vision to build a marketplace. They initially focused on authenticated Funko Pops but pivoted to live commerce after observing customers hacking social media to sell items via live video. The domain "Whatnot" was acquired for half a Bitcoin ($5,000) after the seller initially sent the wrong domain ("WhyNot").
Grant explains that when they started in December 2019, they had limited awareness of the Asian live commerce market (30-40% of all commerce in China) and dismissed QVC as irrelevant. Their insight was that customers don't care about the market structure itself—they care about whether the experience provides value and enjoyment. Users spend approximately 95 minutes daily on the platform, comparable to social media engagement, with over 80% of daily users not making purchases but participating for entertainment and community.
The platform enables small business empowerment, with top sellers generating over $100 million in annual revenue with 20-40%+ EBITDA margins. Whatnot has expanded from collectibles into 100+ categories including fashion (now the largest), fresh food, golf equipment, and more. The company is operational in 10 countries across North America, Europe, Japan, and Australia, with similar dynamics working across all regions.
On trust and safety, Grant reveals that 40% of Whatnot's employees work on this function—roughly equivalent to a police force for a city of 10-20 million people. The company built a rules engine that processes billions of data points per second to detect harassment, shipping delays, refund fraud, and other bad behavior, programmatically enforcing actions using LLMs and automated systems.
Regarding AI deployment, Grant emphasizes preserving the human connection that drives the platform's appeal. AI is used to enhance seller efficiency (metadata inference, analytics, listing tools) rather than replace human interaction. The core experience relies on buyers knowing shop owners and the communities that form around them.
Looking forward, Whatnot's priorities are helping sellers build better businesses, improving the buyer experience through better customer support, reducing shipping costs, and expanding into new categories like cars, liquor, beer, and wine. The fundamental thesis remains unchanged: empower more people to build sustainable businesses while ensuring buyers have an excellent, affordable, engaging experience.
About this episode
a16z General Partner David George is joined by Grant LaFontaine, co-founder of Whatnot, to unpack how a marketplace that started with collectibles evolved into one of the world's leading live shopping platforms. Grant traces the company's origins from selling Pokémon cards online as a kid to discovering live commerce by watching Whatnot's earliest customers hack together sales on social media. They discuss why Whatnot thinks less like a traditional e-commerce marketplace and more like a digital shopping mall, where discovery, entertainment, community, and commerce all happen at once. Today, users spend roughly 95 minutes a day on the platform, and most aren't even buying something on a given day. They also explore how Whatnot is enabling small businesses to reach global audiences, expanding from collectibles into categories like fashion, food, and golf, and using AI to make sellers more efficient without replacing the human connection at the center of the experience.
Key Insights
- Grant sold his first Pokémon card on eBay in third grade for $10, discovering that people would pay for collectibles online—an experience he consciously recreated at Whatnot's founding by focusing on makers and collectors.
- Whatnot users spend ~95 minutes daily on the platform with 80%+ not making purchases, indicating the platform succeeds as entertainment/discovery rather than purely transactional e-commerce.
- Live commerce represents 30-40% of all commerce in China but only single digits in the US, suggesting massive market expansion opportunity if the experience quality improves.
- Grant argues that top Whatnot sellers generate $100M+ annual revenue with 20-40%+ EBITDA margins, proving the platform enables sustainable small business creation that wouldn't be possible on Amazon or eBay.
- Whatnot allocates 40% of total employees to trust and safety operations, framing this function as equivalent to running a police and legal system for a city of 10-20 million people.
- The company built an automated rules engine processing billions of data points per second to detect and programmatically respond to harassment, shipping fraud, and refund abuse without relying on manual review for every case.
- Grant claims that being unfamiliar with established market structure was advantageous because it forced focus on user experience rather than iterating on existing eBay/Facebook Marketplace models.
- The live format solves the trust problem for high-consideration purchases like fresh food by allowing sellers to demonstrate sourcing, explain preparation, and build direct relationships—experiences impossible on static e-commerce.
Topics
Transcript
If you look at e-commerce today, you have to know exactly what you're looking for. Live commerce is over a third of all commerce in Asia. Is that right? 30 to 40% of all commerce is live commerce in China. What is it in the US today? Single digits really. The experience is really broken. Transactions weren't integrated well, shipping wasn't, payments, discovery, CX, and we just thought we could build a better one of those. I want to start all the way back when you were a kid. Your first online sale was a holographic Pokemon card. At that point in time was when like eBay was getting started and I was like, wait, you can make money from…
Full transcript available for MurmurCast members
Sign Up to AccessMore from The a16z Show
How Do You Defend Against AI That Can Hack?
Security teams face unprecedented challenges as AI models become sophisticated enough to autonomously hack systems, escape containment, and bypass traditional defenses. Current cybersecurity tools built to defend against humans and malware are fundamentally inadequate for AI agents, requiring a complete rethinking of defensive strategies.
Ben Horowitz and Travis Kalanick on Building Again
Travis Kalanick discusses his eight-year hiatus from the public eye while building his new company Atoms, focusing on industrial AI across food, mining, and other trillion-dollar industries. He reflects on his evolution as a founder, differences between his Uber and Atoms approaches, and explains his decision not to acquire Lyft, while Ben Horowitz shares insights on entrepreneurship and the changed media landscape.
The Two Ways to Sell AI: Lighthouse or Landgrab?
A16Z partners Joe Schmidt and Andy McCall discuss two competing enterprise AI sales strategies: Lighthouse (targeting high-profile customers to establish credibility in regulated/innovative markets) and LandGrab (pursuing numerous mid-market customers with existing budgets and proven ROI). They argue that early-stage AI founders often mistakenly prioritize prestigious logos over pursuing customers willing to buy, and share lessons from building sales organizations at Meraki and Samsara.
Garry Tan on Taste, Agents and Founder Ambition
Garry Tan discusses the evolution of startup culture, founder ambition, and the role of artificial intelligence in transforming business operations, emphasizing the necessity for founders to pursue their unique insights rather than just following trends.
The CISO Playbook for AI Agents | Datadog
The discussion centers on the risks and opportunities presented by AI in cybersecurity, with an emphasis on understanding and managing malicious intent in code. Datadog's CISO, Emilio Escobar, highlights the importance of proactive security measures, including using AI to evaluate code intent and assessing new AI tools rather than outright blocking them.