On a sun-filled San Francisco afternoon, Peter Szulczewski is climbing the stairs to the top of a Sansome Street skyscraper, past floors filled with Wish data scientists and engineers, pool tables and DJ equipment. Large windows give way to a stunning view of the city. But most of Szulczewski’s customers don’t work in offices like this or live in Northern California coastal enclaves. In fact, most of them don’t have much money at all. Wish’s customers are typically working-class Americans from places like the Florida Panhandle or East Texas, Dollar Store shoppers who find Amazon Prime’s $120 annual membership too rich for their blood.
“Forty-one percent of U.S. households don’t have $400 worth of liquidity,” Szulczewski says, referring to the Fed’s latest estimate. He says customers of his ultra-bargain shopping site have their credit cards declined most often right before payday, rattling off statistics as he easily mounts another flight of stairs, his legs conditioned by decades of weight lifting (he says it helps him relax). The 37-year-old Polish-born former Google engineer is obsessed with ordinary folks’ finances and has used that obsession to tailor an e-commerce marketplace just for them, filled with no-name merchandise shipped directly from Chinese merchants.
Wish was the most downloaded shopping app worldwide in 2018 and is now the third-biggest e-commerce marketplace in the U.S. by sales. Globally, some 90 million people use it at least once a month. Taking a 15% cut of their purchases, Wish doubled its revenue last year, to $1.9 billion. As of its last fundraising round, it was valued at more than $8.7 billion, and Szulczewski’s 18% stake makes him a billionaire. (His cofounder, Danny Zhang, owns just 4.2%.) Szulczewski says investors should expect an IPO in the next year or two.
Wish is hardly the first e-commerce app to track every customer click—Amazon built a $200 billion annual sales colossus on precisely this sort of data. And Wish is competing with both Alibaba’s AliExpress and Amazon in offering shoppers a never-ending stream of Chinese-made items from third-party sellers. But while Amazon keeps adding features like streaming video and two-hour delivery to yoke customers to its Prime membership service, Szulczewski (pronounced sull-CHESS-key) doesn’t worry much about quick delivery or quality. Wish sweaters cost $2 plus $2 shipping, Apple Watch knockoffs go for $9 (plus $3 shipping) and Android smartphones list for $27. Products can take weeks to arrive. Shoppers scroll through an average 600 to 700 items, hypnotized by a pixelated parade of weird and wacky products that scratches the same visual itch as an Instagram feed. Around 80% of Wish’s first-time customers will return to buy a second time.
Shipping is cheap thanks in part to an agreement between China Post and the U.S. Postal Service that allows goods that weigh 4.4 pounds or less to be shipped at low rates. About 15% of Wish’s shipments qualify. It can now be cheaper to ship a package from Beijing to New York than from South Carolina to New York. That’s important because many Chinese merchants can’t afford to use big shipping companies like FedEx or DHL or store their goods in out-of-the-way warehouses as they might with Amazon, says Zhang, who’s based in Shanghai.
Wish is losing about $190 million annually, but it claims it could be profitable if it stopped spending so much on marketing. It has run big campaigns on Pandora and Snapchat, and it’s one of the biggest advertisers on Facebook. In 2017 Wish signed a three-year, $30 million deal to sponsor the NBA’s Los Angeles Lakers, who are popular in China. Los Angeles is also one of Wish’s biggest metropolitan markets by revenue.
Given the competitive landscape, it’s probably money well spent. Mark Zuckerberg recently talked of making it easier to sell stuff on Instagram, meaning Facebook could become a competitor. Amazon’s third-party marketplace accounts for half of all the products it sells (by volume, but not revenue), up from 30% a decade ago. Taking dead aim at Wish, Amazon recently launched a “Bargain Finds” section. Meanwhile, in his efforts to throw heat at China, Donald Trump could upend the deal that makes shipping for some Wish merchants so cheap.
A more existential problem: Lots of the stuff on Wish is trashy, shoddy, even fraudulent. There are hundreds of negative reviews for Wish on review sites like Trustpilot and HighYa. Customers are unhappy about unresponsive customer service, merchants who don’t send their orders and poor-quality goods. Szulczewski has hired Connie Chang, a former community manager at Facebook, to fix the problem. In an act of thrifty opportunism that would impress Jeff Bezos himself, she’s organizing roughly 10,000 Wish users to help the company weed out shady merchants in exchange for free goods and discounts. But Szulczewski seems unfazed by the quality-control challenge, pointing out that sometimes customers themselves are the problem. “We sell 5 million contact lenses a year,” he says. “Someone’s going to sleep in them.”
The man who has 300 million products for sale grew up surrounded by empty store shelves in 1980s communist Poland, on the third floor of a bland, six-story apartment block in Warsaw. He was 11 when the Soviet Union collapsed and his parents moved to Waterloo, Canada, which is about 70 miles west of Toronto and home to an excellent research university. He attended the University of Waterloo, which also counts the founders of Kik Interactive and Instacart as former students. In Szulczewski’s math and computer science classes he met another immigrant, Danny Zhang. The two played soccer together—Zhang was so good he had briefly considered turning pro—and became friends.
In 2004, just before he graduated at 23, Szulczewski started a four-month internship at Google, which then had less than 1,000 employees and was gearing up for its IPO. He lived in a three-bedroom house in Palo Alto with three other interns, churning out code by day and pumping iron at night. “You always got the sense that he didn’t have time for, you know, messing around,” says former Googler Brian Singerman, now a partner at the venture firm Founders Fund and an investor in Wish. Once he became a full-time Google employee, Szulczewski wrote the prototype algorithms for keyword expansion, a feature Google sold to early advertisers to help them expand the number of search terms they could show their ads against. A company selling running shoes, for instance, might forget to ask Google to target searches for “sneakers,” but Szulczewski’s code silently added those keywords so they didn’t have to. Advertisers spent more, and the feature added roughly $100 million to Google’s annual revenue, Szulczewski claims.
In June 2007, Szulczewski transferred to Google’s new office in South Korea and got a lesson in tech outside the Silicon Valley bubble. Koreans preferred search portals that were busy and crowded with information over clean, minimalist home pages like Google’s. Former Google colleague Mark Rabkin remembers debating the point with Szulczewski when he visited. “I could tell he was starting to think like a local,” says Rabkin, who’s now an ad executive at Facebook. The experience would help shape Szulczewski’s strategy for Wish: He learned to focus on building something people wanted rather than what Silicon Valley thought they should want.
Szulczewski left Google in 2009 with enough savings to see him through the next two years. He spent six months at his home computer writing code for software that could predict someone’s interests, such as running or gardening, based on their browsing behavior and match that to a potential product or advertisement, calling the system ContextLogic.
In September 2010, investors put $1.7 million into ContextLogic, the wheels greased by introductions from a friend of Szulczewski’s, Yelp CEO Jeremy Stoppelman. “Their business plan was to go compete with Google on AdWords, in Korea, or some crazy thing,” remembers Jerry Yang, the billionaire cofounder of Yahoo, who invested in Wish through his angel fund, AME Cloud Ventures. But, he adds, “their technology was really good.”
The following May, in 2011, Szulczewski invited his old college friend Zhang, then at YellowPages.com, to join him as a cofounder. They toyed with building an ad business, then had lunch in San Francisco with Yang and the venture capital investor Joe Lonsdale to brainstorm bigger business ideas. E-commerce and mobile kept coming up. More people were using smartphones, but few were shopping on them. Szulczewski became convinced he just needed to apply the “Google methodology” of tracking every interaction, scroll and tap of a customer. Yang was skeptical. Amazon was already 17 years old and well established. “I remember walking out and thinking, ‘Man, this is a hard pivot,’ ” Yang says.
Then things almost ended. In 2011 Facebook got wind of Szulczewski’s recommendation engine and offered $20 million to integrate ContextLogic into its own ad system or, potentially, the software that ranked stories and posts on Facebook’s news feed. When Szulczewski didn’t take the money, one of his investors stormed into the startup’s new office on Pine Street to chew him out. “Go work for someone that’s good at business,” the investor told him.
Instead, Szulczewski kept tinkering. In late 2011 he and Zhang launched Wish’s predecessor, Wishwall.me. Using ads on Facebook, they invited people to come and browse a collection of products curated by Wish. Wish didn’t actually sell any of these products—other sites did that—but people could “wish” for things they wanted and create wish lists, say, for bicycling or home decorating. “Some people are just into that,” says Szulczewski.
As visitors “liked” the bathroom towels and bike speedometers, they earned rewards such as free items and discounts, so they stuck around. Wish amassed tens of thousands of users, and then started matchmaking. It sent emails to dozens of merchants who were actually selling these products on eBay and Amazon. Wish promised to introduce them to a pool of ready buyers, but only if the merchant knocked 10% to 20% off the price they were displaying elsewhere. To prioritize growth early on, Wish didn’t take a commission. If the merchants agreed, Wish pinged its users to say the product they liked was now available for a discount on Wish.
When Hans Tung, an investor with GGV Capital in Menlo Park, California, met Szulczewski in October 2013 he was struck by a heat map. It showed that most of Wish’s sales were coming not from New York and California but from Florida, Texas and the heartland of America. Tung immediately saw parallels with TaoBao, a peer-to-peer marketplace owned by Alibaba that was on track to become the world’s biggest e-commerce site. TaoBao had shown there was an underserved, global market for cheap, unbranded goods but had stayed focused on China. “Wish was TaoBao, on mobile, for the rest of the world,” Tung says.
In 2016 Szulczewski walked away from another bidder, this time his idol. An executive close to Jeff Bezos had reached out and invited Szulczewski to Seattle to meet the Amazon founder. Before the meeting, Szulczewski wrote to Bezos’ executive assistant to clarify that he had no intention of selling his business. “But he still wanted to meet,” Szulczewski remembers. So he took the two-hour flight north for a full day of meetings at Amazon’s headquarters. Citing a nondisclosure agreement that he signed, Szulczewski is reluctant to say much about what went on but says Bezos “wasn’t super-direct” about making an acquisition offer. Amazon’s corporate development attorneys were in the room, though, Szulczewski recalls, as they normally are in acquisition talks. “There was an agenda,” he says. (For its part, Amazon refuses to comment on “speculation.”)
Wish’s open-door policy meant that sales were taking off, but so were quality-control problems—which is somewhat inevitable when you have one million registered merchants, of whom 125,000 are active on Wish each month. By way of context, Amazon hosts an estimated 2.5 million active third-party merchants, while Walmart has roughly 21,000, according to Marketplace Pulse, a New York-based e-commerce intelligence firm.
“I underestimated how challenging [quality control] is,” says Szulczewski. As he walks through the open-plan floor of his San Francisco office, filled with social media staff, he gestures at the symbol of his biggest problem: a collection of cardboard boxes stacked to the ceiling. They’re the site’s most popular products that week, set to be riffled through by staff members for quality checks.
“These merchants will do very interesting things,” Szulczewski says, arching his eyebrows to suggest he has some stories to tell. Here’s one: A merchant once offered tablet computers on Wish for 70 bucks, but customers got a small stand in the mail with a “mostly English” note saying the tablet was coming. It wasn’t. Wish was criticized in press reports last year when makeup on the site was blamed for causing pinkeye. Some merchants have set up more than 1,000 accounts to evade policing. Another trick is to sell a product to great reviews, then source a different, cheaper version and keep showing the old reviews.
To tackle this, Wish automatically culls about 8 million products over a week, or close to 3% of all products available on the site. Most are cut because visitors have scrolled past an item at least 1,000 times and haven’t tapped on it once. Wish also culls products from merchants who’ve received bad reviews or put up fake reviews. To automatically detect the fakes, Wish’s engineers trained its software on reviews that were obviously made up. One example: “Did not expect so soon, the owner of service enthusiasm.” Now the software can identify reviews with similar nonsensical phrases.
Overall, there is a set of roughly 60 rules. If Wish’s algorithms see that a merchant has listed counterfeit products or shipped an order with a fake tracking number, they’ll be “fined” $500. Shipping a package with no product spells a potential $10,000 fine. Wish collects around $3 million a month in fines, and it can do that by simply withholding payments from merchants, says Szulczewski. Merchants can also get kicked off the platform, he adds. On the flip side, good reviews can lead to faster payments or a higher ranking in search.
Zhang, who is 6,000 miles away, overseeing operations in three locations in China with 150 staff members, is just as paranoid about being gamed. “Merchants definitely care about making money first,” Zhang says. The rules and monitoring software don’t always go down well. “The merchants look at Wish’s platform as very robotic,” he says, meaning there’s no way for merchants to negotiate a better deal by cultivating personal relationships with Wish staff.
“It’ll never be perfect,” admits Tung, the GGV investor. “You get what you pay for.” But Szulczewski paints a rosier picture. The flywheel of customer data going into Wish will mean more reviews, smarter targeting software and, eventually, better-quality items. Those unbranded smartphones will get better every year, he says, “just like our iPhones.”