Snaptalent

» see snaptalent.com as it lived, on the wayback machine

~ the obit ~

American college recruiting is an industry with a calendar. Companies send staff to campuses in a season and seniors get hired months before they graduate. A whole layer of software exists to feed that machine. It did not work like that where I took my degree, and Snaptalent's final product was built entirely on top of a process I had no local equivalent for. The distance here is structural rather than temporal.

The original idea was better than the one it died on. Four founders started the company in 2007 and asked a question that still sounds right: why does recruitment advertising live on job boards, when the people you want to hire are spending their time somewhere else entirely? The answer was a distributed ad network. Job ads syndicated out to the blogs and developer sites that engineers actually read, with publishers paid for the candidates they produced. They got into Y Combinator's Winter 2008 batch with it and, by one account, became the first company in that program to book revenue before demo day.

It won recognition quickly, landing on a list of the top ten online recruitment products of 2008. The economics did not follow. The farewell letter is blunt about the mechanism: too few candidate leads per ad impression, so employers stopped buying and publishers stopped getting paid. The network had no way to close that loop. Critical acclaim and an unsustainable cost per hire are a familiar pairing in advertising, and they bought two more swings at the problem.

Both swings landed in the worst hiring market in thirty years. In November 2008 the team pivoted toward recruiting on social networks; in March 2009 it launched a college recruitment platform. By that summer, in their own numbers, roughly 20 percent of American college seniors had a job at graduation, against 80 percent the year before. There was no product to build for a customer who had stopped buying the category.

~ the rap sheet ~

~ three products in nine months ~

The letter they published is unusually self-critical about this, and the self-criticism is not about the recession. In November 2008 the founders had a choice between chasing the short-term opportunity of how companies were starting to recruit on social networks, or committing to a specific audience with a longer horizon. They chose the audience and said afterward it was the wrong call. "Three new product launches in 9 months didn't give us much to build on," they wrote, and the observation underneath it is about morale rather than market size: a team that keeps relaunching loses the trajectory that makes the work feel like it is going somewhere.

The second half of the letter is a tour of why recruitment technology is hard to sell, and it has aged well. Hiring is an enormous market, around ten billion dollars a year in advertising spend at the time, with wealthy customers and very little innovation, and that combination reads to an outsider like an open door. Their explanation for why it stays shut is that HR does not generate revenue, so HR buyers cannot force change through an organization the way sales or marketing can. A vendor that innovates too far ahead of that constraint arrives too early, every time.

~ giving it back ~

The company did not run out of money. It decided the money would do more good back where it came from. "Its been a tough call, as we had enough capital in the bank to last for two years," the letter says, and the decision was to return the majority of the capital raised rather than spend two more years proving a point in a market they had stopped believing in. Hacker News noticed on 17 August 2009, and the thread filled up with people arguing about whether the decision was admirable or premature.

~ the part that reads strangest now ~

Their closing prediction was that the winners in recruitment would be the companies taking money out of job advertising rather than collecting it, through search optimization and referral hiring. That is a reasonable description of what LinkedIn went on to monetize, and of why the job board as a product category has been shrinking ever since. They had the direction right and were holding the wrong end of it.

Snaptalent.com today serves 114 bytes, the generic stub of a domain that someone registered and did nothing with. The farewell is readable only in the Internet Archive, where an October 2009 capture preserves the whole letter signed off by the Snaptalent Team. Everyone involved went on to other things, which the letter more or less promised they would: entrepreneurial hunger intact, lessons taken, the capital already wired back.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »