Gilt Grew Fast but Never Gained Enough Leverage Over Brands

Kevin Ryan’s forecasting approach is to ask who benefits if a company succeeds, who supplies it, and how behavior might change as the trend develops. He says the method still involves instinct and guesswork. As one example, if Shopify does well, the businesses supplying it may benefit too; he points to companies providing data-center “picks and shovels” as another case of looking beyond the most visible winner.

Ryan also describes a missed opportunity. In 2003, he and others discussed falling bandwidth prices. Serving video cost about $10 per thousand, while advertising brought in about $1, so the model did not work. They expected those economics to cross by 2005 or 2006, but did not act on the implication. YouTube started in 2005 and, Ryan says, eventually became a $300 billion company. He offers the example as a reminder to consider what a trend may make possible once its underlying economics change.

Ryan started Gilt Groupe after seeing Vente-Privee, a French flash-sale business he says was doing $1 billion in revenue in 2006. It sold discounted high-end merchandise for limited periods. The idea brought the sample-sale experience online: shoppers who could not get to a New York sale could still buy the brands and goods being offered there.

Ryan remembers seeing about 200 people, mostly women, waiting at a New York sample sale. People in Philadelphia or Austin might want the same merchandise but had no comparable local opportunity. Gilt could reach them, but it was not simply a marketplace connecting shoppers to listings. As Ryan put it, the company had to buy the merchandise, package and sell it, and handle returns. CBS News shopping screens shown alongside his account depict discounted items, product pages with sizes and shipping details, and a cart containing multiple goods.

That operational work accompanied rapid growth.

Ryan says Gilt reached $500 million in revenue in its fourth year. The opportunity rested in part on how brands and retailers were selling at the time. Ryan says Marc Jacobs had no website in 2008 or 2009, when it sold through wholesalers such as Bloomingdale’s. Macy’s, he recalls, had a “terrible” website before improving it to “mediocre” by 2011 or 2012.

The conditions that created Gilt’s opening also changed. Brands such as Marc Jacobs began selling and discounting merchandise directly through their own websites. Department-store websites improved, and other online competitors, including Farfetch, entered the market.

Gilt had been filling a gap, but the gap narrowed. More sellers competed for the same merchandise, and the business became commoditized. Ryan says the company could not find a way around that shift.

For him, the strategic question was whether Gilt could build a moat—enough scale or leverage to remain important to the brands supplying it. It grew substantially, but Ryan gives a concrete measure of the shortfall: Theory once had 20,000 end-of-season items to sell, while Gilt could buy 1,000. That was a meaningful purchase, but not enough to move the needle for the brand. Gilt had become large without becoming indispensable.

Ryan says it was difficult to know in advance how quickly the company would grow or how much influence it would gain with vendors. The business nearly reached the scale needed to shape the market, but did not get there. Being a large participant in a growing category had not given Gilt enough leverage when the category became crowded.

Ryan brought the board a recommendation to sell Gilt when he believed the business was deteriorating. The company had been valued at $1 billion, and a sale for $400 million seemed disappointing. Ryan told the board, “We have a falling knife here.”

Saks ultimately bought Gilt for $250 million. Three years later, Ryan says, Saks offered to sell it back to him for $5 million. The sale was far below the earlier valuation, but in his view it was still better than holding on as the business weakened. The experience left him reluctant to return to e-commerce.

Ryan’s test for a startup is whether it solves a real problem—even one customers may not recognize until a solution appears. He struggles to identify an unmet e-commerce need now: he can get almost anything delivered quickly, return it, and do so inexpensively. By contrast, he sees unsolved problems in areas including cancer, nuclear energy, and psychedelics.