Lessons Learned From DARPA Balloon Challenge

from the it's-amazing-what-people-can-do dept

By now, you've probably heard that a team from MIT won the DARPA balloon challenge, whereby DARPA put 10 red balloons in the air around the country and wanted to see what people could do to find all 10 balloons. The rules were pretty loose, and the team at MIT took all of nine hours or so to locate all ten balloons, through an interesting "crowdsourcing" method. They basically quickly set themselves up as a clearinghouse of information, and made it easily shareable across different social networking platforms, and employed something of an affiliate program to encourage people to get their friends to sign up with the MIT team as well. If you signed up people who helped find the balloons, you got some of the prize money according to your friend network, and so on down through the social pyramid. The team claims that what was most important was the recursive nature of the pyramid, which gave people incentive to participate, even if they knew they couldn't find the balloons.

While some other DARPA challenges, like the autonomous vehicle challenge (to get a totally driverless vehicle to drive a few hundred miles with no help), are cool but seem limited in terms of application outside of the core area it was built for, this one actually does seem to hold a lot of useful lessons that can be picked up on right away, and which can be applied across a lot of different business, policy, IT, public good and many other areas. Some of the key elements:
  • Recognize that there's power in numbers: Recognize that for certain projects, you need a lot of different minds (and eyes) working on things, and that certain tasks shouldn't just be done by "the one best" individual.
  • Make it easy for more people to participate: Once you realize that you need a lot of people, you need to make it easy for them to participate.
  • Give people multiple reasons to participate: Different people have different motivations. Some people just want to belong to a successful project or a leading team to bask in the glow. Others need additional types of incentive. The MIT team offered up monetary compensation in addition to recognition for participation.
  • Give people a reason to get others involved: Sort of a corollary to recognizing the power in numbers, the MIT team worked hard to give people incentive not just to participate and to promote their participation, but also to recruit others to the team as well. This even made it so those who couldn't help finding the balloons directly could still participate in better finding the people who could find the balloons.
  • Align incentives properly: Make sure that everyone is driving towards the same goal, and that the incentives work on top of one another to all push towards that same goal.
  • Look beyond your immediate "group": One of the coolest things I thought about the MIT group was that there was nothing in there that limited it to MIT or the folks at MIT. They immediately recognized that it made the most sense to reach out to folks beyond their immediate circle, which is what helped them get the people they needed involved quickly.
Now, a lot of these may sound obvious, but it's often important to remind yourself of these basic concepts, and it's impressive to see how well (and how fast) these worked in the case of the MIT balloon team. I could see these lessons being applied in a lot of other areas as well. There is a separate issue that the team hasn't discussed yet, but promises to eventually: which is that it also had to deal with a number of bogus entries -- including at least some from a competing team trying to throw the MIT team off the scent. Finding out how they got around such problems would also be quite interesting in terms of better managing these sorts of group efforts.


Reader Comments (rss)

(Flattened / Threaded)

  1.  
    identicon
    Anonymous Coward, Dec 9th, 2009 @ 6:15pm

    a lot of these may sound obvious

    Sound? How about "are"?

     

    reply to this | link to this | view in thread ]

  2.  
    icon
    Steve R. (profile), Dec 9th, 2009 @ 6:20pm

    An Experiment in Finding The Needle in a Haystack

    Based on some other comments that I read at "Freedom to Tinker", this could have been a human behavior experiment for designing a method of finding something that is well hidden. Some of those commenting made references to the search for Bin Laden. I especially like the comment by Anyonmous concerning what may be an incorrect "reward" incentive on our part for finding Bin Laden that: "Chances are, had they offered a couple goats or a camel for good info they would have gotten a lot farther."

     

    reply to this | link to this | view in thread ]

  3.  
    icon
    Marcus Carab (profile), Dec 9th, 2009 @ 6:25pm

    Re: An Experiment in Finding The Needle in a Haystack

    I doubt that the United States has spent this long hunting for Bin Laden without trying to "bribe the locals"

     

    reply to this | link to this | view in thread ]

  4.  
    identicon
    Anonymous Coward, Dec 9th, 2009 @ 6:26pm

    Re: An Experiment in Finding The Needle in a Haystack

    They offered stacks of cash that would make anyone in that area rich enough to buy everything their eye could see and more, and nothing happened.

     

    reply to this | link to this | view in thread ]

  5.  
    identicon
    Anonymous Coward, Dec 9th, 2009 @ 7:01pm

    While some other DARPA challenges, like the autonomous vehicle challenge (to get a totally driverless vehicle to drive a few hundred miles with no help), are cool but seem limited in terms of application outside of the core area it was built for, this one actually does seem to hold a lot of useful lessons that can be picked up on right away

    you must be an IT guy or very closed minded- definitely not an engineer

     

    reply to this | link to this | view in thread ]

  6.  
    identicon
    Thierry Hubert, Dec 9th, 2009 @ 7:02pm

    MIT's excuse of a victory: DARPA's Red Balloons

    MIT acted as a broker to the DARPA project by dividing the winnings according to a race-reward-sharing model for anyone who delivered the coordinates to MIT. They acted more like HBS than MIT on this one. In my opinion this test demonstrates that the Web and social networks are not ready to rival institutions who can basically shortcut and inject themselves as brokers. Additionally, very little was learned or demonstrated as MIT hijacked to potential innovations that could have to emerged from the challenge. MIT won by bribing instead of innovating. MIT should not exhibit confidence in the victory. MIT's victory did not teach DARPA was that wars can be won by trade and economics; the history of conflicts have thought us that very well. I am disappointed in MIT's approach and feel for all who tried to use new technologies to discover new ways to gain awareness.

     

    reply to this | link to this | view in thread ]

  7.  
    identicon
    romy v, Dec 9th, 2009 @ 7:16pm

    same elements also apply to the recent Fat Cyclist and Johan Bruyneel events

    One could analyze the recent events between Fat Cyclist and Johan Bruyneel (http://bit.ly/7yNhF1) and find the same elements being key there as well.

    They are the same reasons why Fat Cyclist was able to mobilize a large number of people and reach the goals that were put in front of him as a challenge. It all started from an innocent letter "hey, wouldn't it be great to ..." - that turned into a great success, not only for Fatty but also for the two charities involved.

     

    reply to this | link to this | view in thread ]

  8.  
    identicon
    Anonymous Coward, Dec 9th, 2009 @ 7:53pm

    Re: Re: An Experiment in Finding The Needle in a Haystack

    Wow. It'd look pretty stupid to STILL be looking for him, ya know...

     

    reply to this | link to this | view in thread ]

  9.  
    icon
    Marcus Carab (profile), Dec 9th, 2009 @ 8:27pm

    Re: Re: Re: An Experiment in Finding The Needle in a Haystack

    pardon?

     

    reply to this | link to this | view in thread ]

  10.  
    identicon
    JP Werlin, Dec 9th, 2009 @ 8:53pm

    The Day Google Didn't Matter

    Mike,

    I agree with many of your points above, however I think you missed a big one regarding Google. I shared my thoughts on the matter here in case you are interested. http://bit.ly/6xTukt

    Great job on Saturday and maybe we will cross paths again.

     

    reply to this | link to this | view in thread ]

  11.  
    identicon
    Anonymous Coward, Dec 9th, 2009 @ 9:03pm

    Re: An Experiment in Finding The Needle in a Haystack

    I doubt they really wanted to find him

     

    reply to this | link to this | view in thread ]

  12.  
    identicon
    fishbane, Dec 9th, 2009 @ 10:36pm

    Shouldn't the lesson be,

    Pyramid schemes are actually good for something afterall?

     

    reply to this | link to this | view in thread ]

  13.  
    identicon
    mryakima, Dec 9th, 2009 @ 11:49pm

    Now "they" have found the tools to stop public dissent during a crisis

    So, imagine a day possibly not very far in the future, when "they" are attempting to crack down on the dissenters (Yes, that could be you and me). With the assistance of a bunch of unwitting helpers this week the powers that be have just been gifted exactly what they need to know to shut down communications among the citizens of this or any other country. This information will also be very helpful to the dictators in Iran.

     

    reply to this | link to this | view in thread ]

  14.  
    identicon
    ., Dec 10th, 2009 @ 3:03am

    Re: Now "they" have found the tools to stop public dissent during a crisis

    Or you can use it to spot corruption too.

    Or you can use it to make the system more democratic with direct input from the people.

    Or you could better coordinate disaster response using people as a network.

    Well the possibilities are numerous the real may be "Does it work all the time"? Which fatctors contribute to success and which don't?

    When people are willing to help it seems to work great but what if it is something unpopular would it work?

     

    reply to this | link to this | view in thread ]

  15.  
    icon
    Steve R. (profile), Dec 10th, 2009 @ 5:16am

    Re: Re: An Experiment in Finding The Needle in a Haystack

    Correct. There is also the political question of whether we really want to find him or not.

    Assuming that we do really want to find him and given the fact that we have not found him we need to reassess our "finding algorithm". Clearly time for Plan B.

    Of course, there is nothing obvious as pointing to a "search for Bin Laden". There could well be another reason for this social experiment.

     

    reply to this | link to this | view in thread ]

  16.  
    icon
    Hephaestus (profile), Dec 10th, 2009 @ 6:24am

    Neat all the things plus one I have

    Most of them were obvious the one I missed ....

    288 note/entry) Add "Look beyond your immediate "group"" to the list of motivations to use list.

    compare the above list to the motivations to use list to see if any others were missed.

     

    reply to this | link to this | view in thread ]

  17.  
    identicon
    Evan, Dec 10th, 2009 @ 6:51am

    The balloon challenge is no where near as innovative as the Urban/Grand Challenge, you should not compare the two.

    All of your lessons are already known, especially to any forward looking social media user. Combing the ages long concept of multi-level marketing with social media is hardly novel. Besides, they won because Slashdot decided to allow their comment to be posted- how is that not favoritism? Do we even know that the Slashdot moderators aren't in for a cut- shouldn't they be? They DID recruit people... conflict of interest if you believe in any sort of objectivity from the Slashdot moderators.

    So I think Slashdot acted improperly, and the other Grand Challenges are FAR more innovative.

     

    reply to this | link to this | view in thread ]

  18.  
    icon
    chris (profile), Dec 10th, 2009 @ 8:29am

    Re:

    Do we even know that the Slashdot moderators aren't in for a cut- shouldn't they be? They DID recruit people... conflict of interest if you believe in any sort of objectivity from the Slashdot moderators.

    slashdot is moderated largely by people selected at random based on their karmic score. moderators get 5 points to moderate posts with and you can only add 1 point to a comment. moderations are then moderated again by an even larger pool of randomly selected individuals in a process called meta-moderation.

    sure it's still possible to game the system, but you would need a large number of people who have mod points available, making a paid system very difficult.

     

    reply to this | link to this | view in thread ]

  19.  
    icon
    nasch (profile), Dec 10th, 2009 @ 8:56am

    Re: Re:

    Perhaps Evan meant Slashdot editors.

     

    reply to this | link to this | view in thread ]

  20.  
    icon
    Eric C (profile), Dec 10th, 2009 @ 10:44am

    Re: MIT's excuse of a victory: DARPA's Red Balloons

    Everything in this post is unbearably stupid. MIT won by bribing? There was a $40,000 reward. If anything, DARPA won by bribing. I read the official rules; they were very loose, and the whole idea was "by any means necessary." They learned exactly what they set out to learn: the quickest way that $40,000 will get you information. Turns out, it isn't Twitter or Wave or Facebook, it's a single, organized hub around which all those other tools can function freely in their own chaotic ways. DARPA knows what the hell they are doing, and if they had wanted things to go a different way, they would have made the rules more strict. The MIT team didn't break the rules or stifle innovation, they simply executed their own idea according to the structure set forth, and theirs ended up being the best way. Simple as that.

     

    reply to this | link to this | view in thread ]

  21.  
    identicon
    Gene Cavanaugh, Dec 10th, 2009 @ 12:02pm

    MIT experiment

    Autonomous vehicle: very important, first step in a very promising direction; BROAD applicability. Breakthrough.

    Crowd sourcing: cute. Been done before, very little real applicability or value, but a sort of fun thing.

     

    reply to this | link to this | view in thread ]

  22.  
    identicon
    Anonymous Coward, Dec 30th, 2009 @ 12:40pm

    You engineers always think things are cooler than people.

     

    reply to this | link to this | view in thread ]

  23.  
    identicon
    Anonymous Coward, Dec 30th, 2009 @ 12:40pm

    You engineers always think things are cooler than people.

     

    reply to this | link to this | view in thread ]


Add Your Comment

Have a Techdirt Account? Sign in now. Want one? Register here
Get Techdirt’s Daily Email
Save me a cookie
  • Note: A CRLF will be replaced by a break tag (<br>), all other allowable HTML will remain intact
  • Allowed HTML Tags: <b> <i> <a> <em> <br> <strong> <blockquote> <hr> <tt>
Follow Techdirt
A word from our sponsors...
Essential Reading
Techdirt Reading List
Techdirt Insider Chat
A word from our sponsors...
Recent Stories
A word from our sponsors...

Close

Email This