I’ve been wanting to give a more in-depth treatment to the T-Shirt company Threadless.com ever since I first encountered them in the course of researching my Wired article on crowdsourcing. Threadless didn’t quite fit within the original parameters of how we were defining crowdsourcing at the time, but under my expanded definition they comprise an almost pure expression of the model. For those of you who aren't part of Generation C, or spending your free time cruising Myspace, Threadless is a perpetual, online T-Shirt design competition. Artists submit their designs; users vote on them; the highest-rated designs are printed and sold back to the community. Simple. Brilliant. Most importantly: Ridiculously cost-effective. When I talked to him this morning, Threadless Creative Director Jeffrey Kalmikoff told me the company is selling 60,000 T-Shirts a month, has a profit margin of 35 percent and is on track to gross $18 million in 2006. This, for a company with fewer than 20 employees. Crowdsourcing can be very good business indeed.
I have two excuses to look more deeply at Threadless’ model today. The first is provided by a post on (Wired editor-in-chief) Chris Anderson’s Long Tail blog on scalability. The general thesis, which fits neatly into his broader ideas on the Long Tail, is that digital businesses can serve niche-markets as easily as they serve mass-markets. Chris rightly points out that crowdsourcing has the potential to be a crucial factor in a company’s ability to do so. I thought it would be interesting to examine Threadless in this regard: Because it “employs” a workforce of thousands, the cost of procuring additional creative is nominal (winning designers receive $2000, and sacrifice all rights to their design in the process). Over roughly five years Threadless has acquired 500 designs on their virtual shelf, about 15 percent of which have been reprinted in response to demand within Threadless' 350,000-strong user community.
Here's where it gets interesting: Threadless has a bulk deal with their printer, which means that it doesn't pay a premium for small print orders. The per-unit cost of producing 15 shirts is the same as 1,500. They could theoretically create a significant business servicing those 15 customers that really want to bring the "cowmouflage" design out of retirement. However, according to Kalmikoff, Threadless only prints orders of 1,200 as a matter of principle: "The whole point is that the community" – as a whole – "determines what's sold." On one hand, it's just this aspect of Threadless' model that makes them such a pure example of crowdsourcing: They're not just generating their core product via crowd labor, but they're using the crowd to determine what product lines are sold, a la Rule 5 from my article. Admirable in democratic principle and elegant in execution, it also raises the specter of the tyranny of the crowd. (Ever notice how many pages of dumb pet tricks and scantily-clad teens one must wade through on YouTube's most popular section to find a watchable video?) And it also illustrates how crowdsourcing can work against Long Tail business models.
On the other side of the scalability question, the company has turned down offers from Target and Urban Outfitters that would have resulted in instant growth for the company. "We decided we were already infinitely scalable," says Kalmikoff, who notes that 99 percent of their sales are through their Web site. "We didn't need a brick-and-mortar presence." And they didn't. If their current rate of growth continues, their '06 revenues will represent a 300 percent growth over those from last year.
My second excuse for writing about Threadless is to clarify the ways in which "the crowd" effectively segments into more specialized aggregations of people depending on the crowdsourcing application. Yesterday I taped an interview for the nationally syndicated public radio show "Here & Now." Joining me was Dr. Alpheus Bingham, the President and CEO of InnoCentive, the network of scientists that played a prominent role in my Wired story. He pointed out to the host that characterizing the 90,000 members of InnoCentive's community as "hobbyists" is inaccurate. Many are highly accredited scientists working at world-class institutions (though others do in fact work from their garage.) Likewise, some of the most successful Threadless designers are duly employed graphic artists with academic training in their craft. I thought it'd be instructive to point people interested in this issue (and it's been coming up a lot) to the designer interviews on Threadless.com's site.
Perhaps it's unwise of me to problematize my own blog's tag line – "The Rise of the Amateur." But what the interviews clarify is that both InnoCentive and Threadless networks create a level-playing field on which amateurs and professionals compete on the basis of their merits, not pedigrees. As the Threadless interviews bear out, the amateur often outperforms the professional.