How Whatnot Built a Global Marketplace

Whatnot co-founder Grant LaFontaine built the company after careers at YouTube on video and at Facebook on Marketplace. The idea surfaced in January 2019 when Grant and co-founder Logan bought 30 to 40 domains during a bar conversation in Tokyo, but nothing moved until October 2019, when they noticed friends returning to collecting Funko Pops, comic books, and sports cards. Both quit their jobs in December 2019 and launched an authenticated Funko Pop marketplace just before COVID. The domain whatnot.com cost roughly half a Bitcoin, approximately 5,000 dollars at the time, and the seller initially transferred the wrong domain, taking 36 hours to correct.

LaFontaine's core diagnosis was that existing marketplaces had broken experiences across discovery, payments, shipping, and customer service, while sellers were already hacking social media by selling over live video without a proper transaction layer. Logan built the first version of the product in a few weeks to close that gap. The company's original name, Always Legit, was rejected by their first designer, and Whatnot was chosen from a one-hour brainstorming session that produced a shortlist of ten names.

Whatnot crossed 1 billion dollars in sales last year and has already surpassed 1 billion dollars in sales so far this year. The largest sellers on the platform are doing upwards of 100 million dollars in revenue with EBITDA margins of 20 to 40 plus percent. Users spend approximately 95 minutes per day on the platform, and on any given day more than 80 percent of users make no purchase, treating the experience as entertainment rather than transactional shopping.

Live commerce represents 30 to 40 percent of all commerce in China and is currently in the single digits as a share of US commerce. The US e-commerce market is approximately 1.2 trillion dollars, and live commerce in Asia is approximately 1 trillion dollars. LaFontaine argues live commerce is demand expansionary, meaning buyers discover things they did not know they wanted, whereas traditional e-commerce largely moves offline sales online without expanding the total market. eBay, which also started in collectibles, was roughly 15 percent of overall US e-commerce a decade ago and is now less than 1 percent, though it still does approximately 80 billion dollars in gross merchandise volume. LaFontaine treats that figure as a floor for the addressable market Whatnot is targeting, with the gap to 1.2 trillion dollars representing the broader opportunity.

Whatnot has expanded from collectibles into more than 100 categories. Women's fashion is now the largest category, dominated by discounted brand-name inventory resembling a TJ Maxx model. Golf is growing hundreds of percent year over year. Fresh food and drink is an early-stage category, including a seafood distributor who sold between 3,000 and 4,000 dollars of fish sourced from a San Diego dock during a pilot. LaFontaine noted that live video and visible seller identity significantly increase buyer trust for high-risk purchases like fresh fish compared to static websites. Cars and alcohol, including liquor, beer, and wine, are identified as future category expansions, with live video seen as well suited to letting buyers inspect a vehicle fully before purchasing.

Trust and safety accounts for approximately 40 percent of Whatnot employees. The platform runs a rules engine that processes billions of data points in under one second to detect bad behavior, evaluating signals such as late shipments and high refund rates to automatically warn, suspend, or ban sellers. A large language model detects harassment in live streams. The live video format itself functions as a structural deterrent to fraud because every transaction is recorded and publicly visible. The team internally compares the operational scale to policing a city twice the size of New York.

On AI, Whatnot uses large language models to infer metadata when sellers configure live streams, reducing manual setup, and provides sellers with analytics that surface where their business is and is not performing. The company has explicitly decided against building avatar tools on the grounds that human connection is the foundational value of the platform. Stated priorities for the next one to three years include stronger seller business tools, an improved buyer experience, lower shipping costs, and potential expansion beyond the roughly 10 countries where the platform currently operates, including the US, Canada, five European countries, Japan, and Australia.