How Whatnot Made Online Shopping Fun Again

Summary

• Grant LaFontaine’s core market call: live commerce is 30-40% of all commerce in China — a “trillion dollar plus market” — versus “single digits” in the US, and the gap is “clearly going to close.” The mechanism is supply-side economics: a fashion brand can pay millions in rent for a Fifth Avenue store restricted to a 9-to-5 local audience, or on Whatnot “set up for $0 and you can have an entirely global audience that is unbounded.”
• David George’s rebuttal to the bear case (“eBay is only so big”): eBay still does $80B of volume but fell from ~15% of e-commerce ten years ago to under 1% today because the market grew past it. eBay is a proof point for at least an $80B market of long-tail commerce to bring online; “the real thing is how much of the gap between 80 billion and 1.2 trillion do you go capture.” LaFontaine goes further — live is “demand expansionary,” more akin to “the inception of department stores or shopping malls” than to e-commerce, which “all it’s really done is take a bunch of offline sales and bring them online.”
• Engagement runs at social-platform levels: users spend about 95 minutes a day on Whatnot, and on any given day more than 80% of them buy nothing. LaFontaine’s buyer value prop in one line: “it’s a fun experience to shop and there’s not a whole lot of fun experiences shopping online.”
• Scale and take-rate posture: over $8B in sales last year and over $8B so far this year, with a deliberately small cut — “we’re not raising fees.” The biggest sellers are “doing upwards of $100 million in revenue” with “very strong EBITDA margins, 20, 30, 40-plus percent,” an inversion of e-commerce platforms that “obscure who you’re buying from.”
• Category and geographic proof points are stacking: women’s fashion is now Whatnot’s biggest category (“almost TJ Maxx-like” discounted inventory), golf is growing “hundreds of percent year-over-year,” and the platform is live in about 10 countries. The counterintuitive learning abroad: dynamics are “incredibly similar” everywhere, making him “incredibly bullish that live is an incredibly durable medium in which to shop.”
• Trust and safety absorbs about 40% of employees, backed by a rules engine ingesting “billions of data points… in under 1 second.” An LLM detects live-stream harassment; the engine programmatically warns, suspends, or bans on late shipping and refund signals. His analog: policing “two New York Cities” — with live video itself as a deterrent.
• On AI, LaFontaine says there is no current need for avatars: if someone goes over the top building one, “people aren’t going to come back to them” because human connection is central. AI goes into listings, metadata inference, and seller analytics; the roadmap is cars, eventually liquor/beer/wine, more countries, and relentless seller tooling.

Deep dive

1. A $10 Chansey, a Tokyo bar, and a half-Bitcoin domain

• LaFontaine’s first sale came in third or fourth grade, in the eBay/Yahoo Auctions era: a holographic Chansey card for $10. George described the buyer mailing a money order; LaFontaine remembers getting it, asking his dad to take him to the post office, and shipping the card. “That’s probably the true origin story of Whatnot” — and his résumé came full circle: Pokémon cards, video at YouTube, marketplaces at Facebook. “Sort of the amalgamation of all those things.”
• The founding had two stories, both true: drinks in a Tokyo bar with Logan around January 2019 and 30 or 40 domains bought on the spot; elsewhere, LaFontaine refers to those as 100 domains. Then months of nothing. The real trigger, sometime around October 2019: friends re-collecting Funko Pops, comics, and sports cards — “everyone like us was older, they were starting to have some disposable income, and they grew up on a mobile phone… if someone was going to figure it out, maybe it could be us.” Both quit in December 2019 to launch an authenticated Funko Pop marketplace.
• The name was originally “Always Legit” — LaFontaine’s placeholder logic: “Who cares what it’s called if you got nothing?” After a forced hour-long brainstorm, Logan reached Whatnot’s owner’s representative, who said “I’ll sell it to you for a Bitcoin”; Logan haggled to half a Bitcoin (~$5,000 with BTC at $10K), the seller first transferred Why Not instead of Whatnot, and 36 hours later they had Whatnot.

2. “Users don’t give a— about the market” — ignorance as advantage

• They knew nothing about Asian live shopping and dismissed QVC (“no one my age wants to shop on QVC”). George’s read — that market ignorance was an asset, since experts just iterate on what exists — gets LaFontaine’s strongest agreement of the episode: “1,000% an advantage.”
• George supplied the framing that “users don’t give a— about the market”; LaFontaine agreed with the underlying point: consumers care whether they will enjoy the experience and whether it provides value. Chasing the Asia analog as “a market,” as competitors did, distracts from building a great user experience.
• The actual pivot came from watching their Funko customers “hacking social media” by selling in live video with everything broken — “transactions weren’t integrated well, shipping, payments, discovery, CX — and we just thought we could build a better one of those.” Logan locked himself in his room for a few weeks; the first live show sold out, sold a lot of items, “and the rest is history.”
• Product framing worth keeping: Whatnot is “analogous to visiting a store” — “maybe you know you want a thing, but you don’t know what that exact thing is” — versus legacy e-commerce, where “you have to know exactly what you’re looking for.”

3. The sizing debate: China’s 30-40% versus America’s single digits

• George’s answer to the perennial knock (“Etsy is only so big, eBay is only so big”): eBay was ~15% of the overall market roughly ten years ago, still does $80B of volume, and is now under 1% — the market simply grew beyond it. eBay is a proof point for at least an $80B market of long-tail commerce to bring online; “the real thing is how much of the gap between 80 billion and 1.2 trillion do you capture.”
• LaFontaine tears the eBay comp apart differently: live enables “a whole swath of different types of sellers” — small businesses and brands with brick-and-mortar or Shopify stores that would never list on eBay — and live is “demand expansionary,” since e-commerce so far has only “taken a bunch of offline sales and brought them online in a more efficient way.” Both land on the same analog: “the inception of department stores or shopping malls,” which expanded demand rather than shifting it.
• Why the US gap closes, in his telling: millions in Fifth Avenue rent for whoever’s in New York 9-to-5, versus $0 setup and an unbounded global audience. “I don’t think you’re going to be able to ignore it” — that’s why China sits at 30-40%.

4. Entertainment-grade engagement funding small-business economics

• Users spend about 95 minutes a day — social-platform territory — and “more than 80% of people don’t purchase on a given day. They’re just there watching, having fun.” The mall-as-hangout analogy carries it: many trips ended without a purchase.
• The seller thesis inverts 20-30 years of e-commerce, which “got so efficient” partly by “obscuring who you’re buying from”; Whatnot is “the polar opposite” — your store, your brand, your returning customers — “generating billions and billions and billions of dollars for small businesses,” from one- or two-person operations to a couple hundred people at the high end.
• The numbers: over $8B in sales last year, over $8B so far this year, with the biggest sellers “doing upwards of $100 million in revenue” at “very strong EBITDA margins, 20, 30, 40-plus percent.” Whatnot takes a small fraction and isn’t raising fees — “it’s really important that the businesses take the majority of that.”

5. Category and geographic proof: fashion is now #1, and Europe rhymes with the US

• Collectibles are “going mainstream” — celebrities and athletes in, and “a lot of the new art that people are collecting is now trading cards going for millions of dollars”: the “one-of-one LeBron James refractor” is an example of what people now buy where prior generations might have bought famous artists.
• Women’s fashion is now the biggest category, dominated by “almost TJ Maxx-like” old-season inventory at deep discounts (the DSW analogy for sneakers); live moves enough volume that sellers can afford the markdowns. Golf is “growing hundreds of percent year-over-year”; fresh food is early but “anything that someone would build a retail shop around would work on Whatnot.”
• Best specimen of the discovery thesis: a seafood distributor messaged LaFontaine on LinkedIn, went live the following week, and sold $3,000-$4,000 of San Diego fish in the pilot. Video solves the trust problem in a category people fear buying online — you see the fisherman, ask how he packs it; “if you’re on a static website — well, I hope it’s good.”
• About 10 countries now: US and Canada, five in Europe, Japan, and Australia — and the counterintuitive learning is that “the dynamics are incredibly similar” everywhere. “We’re incredibly bullish that live is an incredibly durable medium in which to shop.”

6. Policing two New York Cities — and why AI stays backstage

• Trust and safety is “about 40% of employees,” and “the punch line on trust is measure everything.” The rules engine takes in “billions of data points… in under 1 second”: an LLM detects live-stream harassment, and the engine programmatically warns, suspends, or bans based on seller history, late shipping, or high refund rates — with weightings that “will always change.”
• His governing analog: tens of millions of users make the T&S team “the police force and the legal system” of roughly “two New York Cities” — imperfectly run by definition, but with far fewer problems than the real thing, helped by the format itself: “everything’s on video… there’s not much you’re going to get away with on live video.”
• On consumer-facing AI, the platform self-reinforces against loss of human connection: “if someone goes over the top… building an AI avatar, people aren’t going to come back to them.” AI’s job is keeping it human while running the business efficiently — listing items, inferring live-stream metadata for discovery, and providing seller analytics. “Today we don’t see a need for avatars because that’s not why people come to Whatnot.”
• The 12-36-month roadmap is deliberately simple: seller tools, better buyer experience (customer support, lower shipping prices, selection), some more countries — plus cars (“you should be able to buy a car on Whatnot… might not be good for my bank account”) and eventually liquor, beer, and wine.