Eighteen months ago, we had a big idea. We thought it was crazy that people make career decisions with so little information about the industry, company, or job they're getting into. So we set out to build something called LifeSwap, a peer-to-peer marketplace for job shadowing, where people could "step into someone's shoes" and see what a job is actually like before committing to it.
It didn’t go according to plan.
We knew we weren’t recruiting experts, so rather than building an elaborate marketplace, we threw up a WordPress site with nice looking job listings at unspecified locations. While we certainly could not post this achievement to our GitHub profiles, it was an extremely efficient prototyping tool. More on this in future posts.
We wanted to test our idea before barreling down a road lined with presumption and subjectivity. We were following Eric Ries’ mantra of being scrappy—using limited resources (which for us meant time) to refine our hypotheses: What should we be building, and for whom? We weren’t recruiting experts, so we had to listen to our early adopters to figure it out.
We went WAY deep into eating our own dogfood. In fact, we were our own first users, hosting a large variety of different swaps. In fact, we hosted our own first swap (Bastiaan and Mike swapped apartments for one night), our most expensive swap (a visit to our d.school class with Michael Dearing), and one of our most popular swaps (hosting people at LifeSwap – you guessed it, to see what it’s like to work at a startup).
People clicked most on job listings for "farmer" and "baker," so those were the businesses we went after first. We cold-called and then visited a farm and a bakery, both here in northern California, to talk to the people there and get them excited about our service. We found people who love their jobs and want to share their openings with like-minded people.
We thought we were savvy—we didn’t outsource the arduous task of visiting potential hosts, accompanying guests on their swaps, taking photos, interviewing all participants. We wanted to do it all ourselves. We learned fast; if we hadn't been so scrappy, we might have spent more time and money building a site catering to white-collar jobs, thus missing the mark on what our early adopters actually wanted.
As we added more and more companies (whom we dubbed "hosts") to our site, our offerings started to become increasingly disparate in scope. We were listing hosts ranging from the culinary arts to artisanal glass-blowing to service sector, technical and highly-specialized shadowing gigs in PR, startup web development, and entrepreneurship. Very soon we realized that there were too many directions in which we could run. Should we find another 10 bakers? Or an investment banker and painter instead? How could we hone our efforts across disparate job categories?
We decided to double down on features for tech companies—where we were getting significant traction. We had been getting a lot of inbound interest from companies wanting to use LifeSwap as a hiring tool. What better way for them to vet a candidate than during a two-hour "swap"? Tech companies were willing to sign contracts rewarding us substantially for guests who turned out to be hireable. It all sounded great.
What initially seemed like a victory began to slip away. Tech companies loved using LifeSwap to host visitors, but only if they were top Computer Science students. We gave them what they wanted and sent groups of up to eight students to top tech companies including Airbnb, Dropbox, and Pinterest. On the surface, this was a great achievement. We had companies willing to pay and guests interested in visiting these companies.
Unfortunately we observed that many students didn't need these visits to get recruited by top companies, they could make contact themselves. We tried to be really disciplined about knowing what kinds of experiences would comprise a really big, successful marketplace. Marketplaces are a special animal, something we will discuss more later – but we could see that it would be difficult to reproduce these types of matches between tech companies and top engineers in every major city. Without a consistent supply of top-tier candidates AND hosts, our market wouldn’t clear in the longrun, and the economics of LifeSwap would be challenging.
If our primary monetary incentive was to bring in the very best recruits to the top companies, then our day-to-day would, in the longrun, feel more like that of a headhunting business—not exactly in line with our dream of helping people find jobs they're passionate about. Our team had no expertise in headhunting. It wasn’t authentic to us. Where was our edge?
We came together as a team and discussed next steps. It was clear that we were pessimistic about pivoting LifeSwap into a viable business, but we were so committed to building a great product together that none of us just wanted to call it quits. We were still bothered by the same core problem: it was simply too hard for people to access and digest key information about the workings of great companies.
When we reassessed LifeSwap, we noticed that one of the top destinations on our site was our Map feature, which is still live here. From analytics, we could tell people usually spent significant time clicking around company profiles, where they could "vouch" for companies they knew and liked. We had inadvertently created a catalog of fast-growing companies and a graph of those companies' reputations. All this time, while working on our product, we were building a core asset we didn’t even know we had.
At the next Stanford Hackathon, we built on our new revelation by leveraging CrunchBase data to create a visualization of companies' funding relative to their size. Having worked in investing and tech M&A before graduate school, we saw how we might have used a company intelligence tool just like this. It was authentic to our backgrounds and interests—and we were energized!
We quickly discovered other people who were hankering for a product like ours. Current and former classmates of ours needed company intelligence for their job search; journalists needed it for research and content. Could we fill the enormous gap left by Bloomberg and Thomson Reuters when it comes to data on private companies? Private company intelligence was certainly broken, but with more and more data points available online, we figured we could devise new "leading indicators" and measures of company performance and traction, making up for the dearth of financial data available for non public organizations.
Our scrappiness is paying off. We may have fallen serendipitously into a position to disrupt the antiquated business intelligence sector using modern UI, advanced natural language processing, and a customer-centric approach to building our product.
Above all, the problem we are addressing remains an authentic one. We are our own customers!
Bastiaan Janmaat, originally from the Netherlands, worked in Goldman Sachs' London-based venture capital investing arm before earning his MBA at Stanford, where he was also co-president of the Entrepreneur Club.
Alden Timme hails from Stanford's Computer Science program, where he earned his BS and MS with a specialization in AI, NLP, and Statistics. He also contributed as a Stanford Research Assistant.
Mike Dorsey, after working at Southwest Airlines and Cisco, started a profitable lead-gen business before attending Stanford for an MS-MBA dual degree. He founded FounderSoup, an entrepreneurial community at Stanford.
[Image: Flickr user Zeevveez]