While many people think of Upwork (the company produced by the merger of oDesk and Elance) as an outsourcer of telework to low-wage countries, more than half of its work is performed by workers in the U.S. In addition to being one of the premier companies in the on-demand economy, Upwork has a great deal to teach us about on-demand skills assessment, on-demand training, and the benefits and risks of worker monitoring. Gary led oDesk until the merger with Upwork, and has a unique perspective on the on-demand economy.
— Tim O’Reilly
When you start a marketplace business, you really have to figure out your business model, and who you’re serving. In the very beginning we were a staffing firm. We were helping companies to find and to hire and to manage and to pay remote talent, but we were doing it in a very hands-on way. We would screen all of the clients. We would screen all of the talent, and, we would put the two together, but we didn’t have a publicly facing website to do that. It was all what I call hamsterware. So essentially we were a staffing firm helping you to find, hire, manage, and pay talent, but we were doing it in a high touch way. It looks one way, but behind the scenes you have hamsters on wheels running to keep it powered, and those hamsters on wheels were actually people who would hand-match a developer, in the early days, with a client who needed developer work done. All of the matching, and everything on the back end was handled by humans.
The beauty of eBay was that once buyers came and started listing stuff, that brought the sellers. And then once the sellers came, what they realized is sellers could be buyers and buyers could be sellers. And so you had this cross-pollenization where everybody could be every role in the marketplace and that was the beauty of eBay.
In the context of OODA—observe, orient, decide, and act—if the decision is reversible, you should make it quickly and just go. If it’s irreversible, you want to put a little more time and energy into it, but you still want to make quick decisions. We looked at [the decision to become a marketplace rather than a staffing company] as reversible. We could pick one. If it didn’t work out we could pivot in six months.
Let me give you a story. A friend of mine had a website, built on to a disk….I went in and looked at his assignment because I wanted to make sure he had a good result.
He gave his contractor five stars and great feedback. Then I spoke to him. I said, “Hey looks like your assignment went well.” He said, “Yeah it was pretty good.” I said, “Pretty good? You gave the guy five stars and glowing comments.” He goes, “You know, I didn’t want to ruin the guy’s life. He was good. He wasn’t great. It was only $1500. It would have cost me $10,000 locally, and the guy begged me for good feedback.”
“I just didn’t want to screw his future on oDesk.” That was a light bulb for me. I was like, oh my God, I wonder if we have grade inflation? So one of the things that worked really well was private feedback. So this is the public feedback. But why don’t you tell me behind the scenes, what did you really think? Would you hire this guy again? Is there any coaching you would give him?
Then, that data we’ll never show to clients. But we can use that data to figure out who’s really good. So we can use that in our algorithm to showcase people. One other piece that was really valuable is we were showing people with incredibly positive feedback. But the problem is that clients weren’t happy because they could never hire them.
They were never available. Why? Because their feedback was so good, they were getting all the jobs. So in addition to feedback, we had to take into consideration availability and also recency. You may look like you are available because you don’t have any clients right now. Your feedback is exceptional. But you took a permanent job three months ago, and you’re not really active in the marketplace.
So we’ve got to take all of these elements into consideration. Your feedback, your reputation, your test scores, your verifiable work history, your recency, your citizenship, your karma, your eagerness, your availability, and put them all together to really make an effective match between the best client and the best freelancer every time.
Any time we see a demand for skills (where one does not a series make) we have data scientists poring through this data, and we’re trying to see where is the demand going—and if we see significant demand start to emerge, we can create another category or a sub-category.
Gary Swart is a venture partner with Polaris Partners and until 2014 was the CEO of oDesk, the world’s largest online workplace—which has more than 8 million registered contractors, and over $1B in work delivered through the platform. Gary is a thought leader in entrepreneurship; how best to hire and manage teams; and the future of work, including online work. He is passionate about helping small businesses thrive, fueled by his extensive experience working with startups and small businesses and mentoring entrepreneurs and business school students. Gary has spoken at the Inc. Leadership Conference, The Economist’s Ideas Economy panel, South by Southwest, TechCrunch 50, TiECon, GigaOM’s Net:Work Conference, and at Harvard Business School, which teaches a case study on oDesk. His commentary has appeared in a variety of publications including Forbes, TechCrunch, The Washington Post, and The Next Web. He has also appeared on TV and radio shows including the BBC, National Public Radio, CNBC, Bloomberg TV, and Startups Uncensored. Previously, Gary led SMB Sales for the Americas at IBM’s Rational Software Product Group, and also served as VP of worldwide sales and operations at Intellibank. Gary holds a B.S. in business administration from the University of Maryland.
©2015, O'Reilly Media, Inc. • (800) 889-8969 or (707) 827-7019 • Monday-Friday 7:30am-5pm PT • All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners. • firstname.lastname@example.org