Quixey Hackathon #1 Recap
September 13th, 2012 | Posted by in Quixey NewsThis post was written by Liron Shapira, co-founder and CTO.
Thursday, August 9, 2012 was a big day at Quixey – our first company hackathon. At 10am, we dropped everything, split into teams of 4-5 Quixiers, and spent the next 24 hours hacking together prototypes of crazy new ideas.
The hackathon was data-themed, and there were three prize categories: Best Data Collection, Best Data Analysis, and Best Data Visualization. The prize for each winning team was a $100 Github credit for each team member, and naming rights to one of the conference rooms in the new office we just moved into. Thanks to Github for generously donating these prizes in support of the Hackathon!
The winners were decided by our hackathon judge, Dan Appleman. Dan is a serial entrepreneur and programming expert with over a dozen published books (mostly about programming). We asked Dan to judge the projects based on how much value they could potentially add to Quixey.
Hackathons are everywhere in Silicon Valley (it’s a real hackathon-athon here). We wanted the Quixey hackathon to be special, and we wanted everyone to be productive and have a great time. So we did some things differently:
- We did it on company time
If we had done a weekend hackathon, we wouldn’t have had to reschedule our meetings and deadlines. Weekday hackathons are harder and more expensive. So why did we do it on a weekday? Because we wanted to run a hackathon like a serious company effort, not a gimmick. Taking the hackathon seriously meant trading it off against the other priorities in our week.
- We all participated
Most hackathons attract a certain demographic – the single and junior engineer. We thought every Quixier – single or married, intern or VP, engineer or non-engineer – would have something valuable to contribute. So we assigned everyone in the company to a hackathon team.
Sure enough, every employee found something to contribute. Quixiers in Marketing worked on PowerPoint slides for their team’s presentation. Quixiers in Business Development worked on exporting Salesforce data for their team’s project to analyze. And of course, Quixiers in Engineering and Design built prototypes.
Some people went home to their families after their day of hacking; others kept hacking late into the night. I gave my teammates a ride home at 4:30am and picked them up four hours later. When we got back to the office, we saw a few Quixiers who were still there after pulling all-nighters.
- We mixed the teams
After Quixey’s two years of nonstop hiring and exponential growth, things can’t help feeling a little bigger and less personal. Normally, each of us only gets the opportunity to work with a few other Quixiers. So for the hackathon, we assigned everyone to teams that had almost zero overlap with normal working teams.
We were surprised how great it felt to work intensely with talented and passionate new teammates on Thursday. And on Friday, it was amazing to see all nine teams deliver a great presentation. After seeing how well randomly-mixed groups work together at Quixey, we really feel like the whole company is one unified team.
- We reversed leadership roles
A lot of our summer engineering interns were going back to school soon after the hackathon, and we wanted to make their last Quixey experience great. So we declared that the most junior engineer on each team was the captain. That rule left us with a lot of ambiguous captain assignments, and we had to use a nondeterministic algorithm from game theory to break the ties: rock-paper-scissors.
If we hadn’t assigned team captains, we think interns and junior employees would naturally have looked to senior employees for leadership. We were happy to see how reversing the leadership roles successfully created a balanced team dynamic during the hackathon.
- We had a cooperative spirit
The Quixey hackathon was really about exploring potentially great ideas for Quixey. We still wanted it to be a competition, but we also wanted to encourage cooperation among teams. One thing we did was purposely avoid giving extravagant prizes. We wanted people to feel good about taking time to help other teams, not feel like they were missing out on a new Macbook Pro if they didn’t win first place.
A lot of teams ended up needing help. Some people actually finished their own projects and stayed late to help other teams. That spirit of cooperation was one of the main things that made our hackathon a success.
- We encouraged teams to think big
Most hackathons encourage thinking small. It’s tempting to just build a prototype of a simple product because you know you can have a polished demo ready in 24 hours.
We wanted to encourage teams to hack on ideas that mattered, ideas that could have a big impact on the company five years from now. So we announced that we would be judging projects according to their expected value to Quixey – not just how cool or polished the demo looked. In fact, the team that won the prize for Best Data Collection didn’t present a working demo. They presented research, statistics and graphs that made a compelling argument for implementing their idea.
- We connected the hackathon to Quixey’s product road map
When we encouraged the hackathon teams to take risks, we knew a lot of the projects wouldn’t work out. But if they did work out, we wanted them to have a real impact on the direction of the company. So after our day of creative thinking and exploration, we shifted gears and took a hard look at our product road map.
The hackathon projects blew away our expectations. One of the projects was such a success that we incorporated it wholesale as a major upcoming search feature. A few of the others are planned as product features and internal tools. And most of the projects have some value for our research and patenting efforts.
We have to keep all the projects classified, but here’s an overview of our three hackathon winners:
Juggernaut’s project focused on a new kind of signal we could use to parameterize our machine-learned search. It was a great first effort to take our search development in a direction we hadn’t had time to explore before.
Best Data Analysis: Team Tomato Jedi
Tomato Jedi’s project analyzed how we could relate apps to a web user’s intent. Judge Dan awarded them the prize because their project was interesting from both a technical and business perspective.
Best Data Visualization: Team Entropy
Entropy’s project dealt with visualizing the “functional web” – the web of apps and their interrelations. They took a risk by experimenting with some clever graph algorithms, and produced a demo that was both beautiful and productizable. Their presentation shocked us all.
When you have a world-class team, pushing everyone to the limit for 24 hours is a thrill. It feels like driving a new sports car and stepping on the gas. After the hackathon, the most common response I heard was how amazing it was that all the teams pulled off their projects. And it was nice to give our interns something special to remember from their summer at Quixey.
We’re so happy with how Hackathon #1 turned out that we’ve decided to continue with regular hackathons. Stay tuned for Quixey Hackathon #2!
You can follow any responses to this entry through the RSS 2.0 You can leave a response, or trackback.