This was a fun side project I built with Adam Doppelt (co-founder of UrbanSpoon). Pretty interesting data so far-- we're trending to having pretty solid reputation data on hundreds of thousands of folks after our first week. 50% of users do 50 or more duels and 20% do 100 duels or more.
The goal of the project was to get an opportunity to work together a bit, explore some new technology (RackSpace Cloud, MongoDB, beanstalk, bluepill, HAML), and do some fun design work.
How did you get enough folks to sign up (without any prior announcement) to get reputation data for hundreds of thousands of folks? Just from sending it out to friends?
Great work Tony! This is highly addictive and made me informed on what some of my past colleagues are up to. Apparently an old classmate decided to ditch programming and become a police officer. O_o
Fun project. But highly concerning if it gains real traction -- As a general rule, people are only on my LinkedIn profile if I already think they are good to work with. This site makes me turn half of them into losers. That is not a fair assessment.
Also, when I deactivated my account, it said that coworkers would no longer be able to see me. So... people who have never ever seen this site are visible, but ex-users are not? Something feels sketchy there.
Running your coworkers, particularly if they're your friends through a series of forced choice tasks and then displaying the results publicly is probably a good way to hurt feelings and alienation.
Ask yourself whether you'd do this for your facebook friends.
Adam Doppelt (co-creator) here. Feel free to chime in with technical questions and I will attempt to answer them. That's half the fun of a side project!
First, are you seeing good convergence on numbers? How many votes does it take, and how long does it take to acquire them?
Given your background, have you considered building a similar "duel" for restaurants? Mostly I'm curious how the restaurant duel numbers would compare to Urban Spoon's rankings. (There is the problem that not all users will have been to all restaurants, so maybe restrict to a small popular subset in each geography?)
We don't have those kind of detailed stats yet, unfortunately. Right now we're just trying to keep things humming along. It's just me on the tech side!
Resto duel is an interesting idea. Ethan (from Urbanspoon) actually mocked this up at one point but we thought it was kinda lame. The results just weren't that compelling, sadly.
TechCrunch drives a bunch of traffic, generally, especially something like Cubeduel that plays well to that crowd. We're somewhat insulated in that many people lack LinkedIn accounts and bounce off the front door.
The first system to fall over was our message-queue that talks to LinkedIn for logins and refreshes. I had to spin up a separate machine at Rackspace just to help logins along. Hopefully that'll hold things steady for a bit.
Suggestion: ratings should be based on some kind of logistic model (like Elo ratings for chess) instead of the percentage of duels won. e.g., feed the scoring history into Bayeselo or something to get a ranking.
It seems a bit interesting, but I see a couple of points.
1. I have a small pool of people on my linkedIn profile, so I don't think there are a lot of relevance to my results. IE: I will run out of battles very soon
2. How well does this scale to people that have a huge number of people on their profile, for example a recruiter? They probably aren't qualified to give their opinion.
Or, it means that you worked in industries where LinkedIn use still isn't widespread, and/or places where even if people have an account, they don't accept connections.
Ha! A politically correct hot or not. I found it incredibly easy to get started and quite addicting actually. Before I knew it I had voted on over 30 co-workers.
Same here ... very easy to get started ... and kind of addicting ... not "Angry Birds"-addicting, but addicting none-the-less. Get some ads up there and monetize!
That being said, I do wonder if LinkedIn is the right venue. The premise is very similar to something like "hot-or-not," and I'm just not sure that the "professional" crowd on LinkedIn is going to take to it.
Maybe you can consume another social network's API with a younger audience though.
This could be a goldmine for HR and recruiters to help vet someone's likability before sending them on an interview.
Any plans to roll out a ranking badge that people can embed on their sites/resumes/linkedin page that gets dynamically updated as more votes come in for each person?
I'm really sorry. I wanted to show you cubeduel, but it looks like you have cookies disabled. Please turn on cookies and try again!
Try Again"
No Cookies
I'm really sorry. I wanted to take a look at what this was about. It would have been nice to see some sort of description without having to let you mark my computer.
The interface and mechanics of the site looks like an emerging pattern in data collection to understand the relations between nodes on the social graph. Seems like a throwback to the days when psychologists would show physical cards and ask us to choose one.
I know it was just meant to be a little fun thing, but that said it sometimes asks absurd questions like, "Would you rather work with the marketing intern or the CEO?"
I'm not sure that's such an absurd question, depending on the CEO and the intern.
I can right away think of one intern I've worked with that I would really like to work with later on, and I've worked with one CEO that I would never want to work with again... (only one intern and one CEO because I haven't worked with that many of either in fact)
I think this has huge implications, not just in this space, but in others (which restaurant? which vacation? other facets of favoriting co-workers? ways a person could improve) -- ties in very well with Jon Bischke's article on Reputation Graph:
What's the gaming algo you are using to decide who to duel against whom? I'm assuming your not simply randomly displaying pairs from the same company N number of times? Do winners get pitted for more duels then losers?
I like the concept - easy to start playing, and it kind of sucks you in - especially when you start contemplating coworkers that you haven't seen and worked with in ages.
I am having an issue where CubeDuel thinks that I only have 2 connections, and keeps bringing them back. Not sure why this is happening, or if you've seen this before.
That can happen if we can't find good overlap between your work history and the work history of your LinkedIn connections. We make reasonable attempts to normalize the resumes, but we can't account for variations, or even plain ol' typos. Double check your work histories and see if you can line them up better.
Still, wouldn't it be better to not do repeat votes? Easy enough to do a query to see if a duel's already been done between 2 folks, though I wonder if that would cause too much of a load increase on your database for you then?
Hmm, it seems coworkers might be able to infer who voted against them if only one person in their office has voted. What sort of protections do you have against this?
The goal of the project was to get an opportunity to work together a bit, explore some new technology (RackSpace Cloud, MongoDB, beanstalk, bluepill, HAML), and do some fun design work.