Wednesday, September 22, 2010
Collective Problems
From Dumb Crowd to Smart Crowd...
For a long time the popular thought was that crowds were stupid. When you gathered a group of people together they became a mob and began to think as one, a phenomenon commonly called ‘mob mentality’. The more people there is, the dumber they become. Everyone’s individual ideas are squashed into one big idea that everyone consents to, meaning the end result is a mainstream idea.
This all changed with a book by James Surowiecki called ‘The Wisdom Of Crowds’. In it he revolutionised the idea of crowds and showed that they could in fact be smart. A bunch of individuals could in fact produce new, creative and innovative ideas, if the ideas were not ‘averaged out’, but instead nurtured.
...To Collective Mayhem
But this new train of thought was widely adopted and the collective intelligence trend boomed.
“I think a lot of people kind of went too far in the other direction, in saying, 'Wow, it's wonderful if we just let everyone participate on the web like in Wikipedia, everything will turn out great.', says Professor Thomas Malone, the Director of the Centre for Collective Intelligence at the Massachusetts Institute of Technology.
This saw the rise of a plethora of websites jumping in, eager to harness the power of collective intelligence, but not knowing how to go about it successfully. This meant many, many failures.
Trying to be all Organic
You can’t engineer collective intelligence by setting up an artificial situation, yet neither can you let the process be completely organic, and arise all on its own accord.
'There is the misconception that you can sprinkle crowd wisdom on something and things will turn out for the best. That is not true, it's not magic.' (Thomas Malone).
The answer, is that a middle ground approach is needed, as boring as that may be. The most successful types of collective intelligence require planning, rules, organisation and control first and then need to be let go and allow participants to contribute freely. This is a far cry from the sort of ‘melting pot’ mentality, where everyone’s ideas are thrown together and simmer automatically to generate an end product.
Too many Indians, not enough Chiefs
Then it is fair to say for quality collective intelligence to arise, some form of leadership is necessary. No matter how many people you have, nothing will get done if no-one knows what needs to be done. Some direction needs to be provided, otherwise all you gain is a bunch of loosely formed and competing ideas...a big indecipherable mess. James Gosling, the creator of software program ‘Java’ says that it is no coincidence that thousands of companies which use community knowledge are flailing and ineffective, while the company Apple is run by a “dictator with good taste” leads the way in technology design.
Use Me, Don’t Abuse Me
The web-based business ‘Threadless’ is a good example of how to use and not abuse collective intelligence. ‘Threadless’ gets users to submit their own artistic designs, which are then voted upon by other users and the top-rated designs are made into T-shirts which are then sold on the website. Collective Intelligence is about the community on Threadless, voting and discussing t-shirts designed by individuals. Dumbness of Crowds would be expecting the Threadless community to actually design the t-shirts together as a group. The following picture illustrates the 2 common results which occur when groups are asked to design something together, in this case the perfect dog:
Our Experiment
Our ‘Facebook as a way of generating ideas’ experiment fell victim to these common problems:
1. We got too excited too soon. We bought into the myth of collective intelligence as some kind of utopia where we would get a creative result with minimal effort by just ‘putting something out there’.
2. We created an environment that was in some ways too artificial (using a Facebook set up), but in other ways too organic (having little control over it). We modified and modified, but didn’t achieve the perfect balance which is needed to generate quality ideas!
3. We had no real leadership or direction to the collective intelligence.
4. This meant that the results gained were both types of ‘dogs’ from the above image. We got alot of average, mainstream intelligence in the form of viral videos. Whilst amusing, they didn’t generate any new ideas. If we attempted to put these ideas together we would have gotten the Frankenstein Dog - a mismatched collection of ideas with no glue.
For collective intelligence to work it cannot be left to its own devices, it needs work.
For a long time the popular thought was that crowds were stupid. When you gathered a group of people together they became a mob and began to think as one, a phenomenon commonly called ‘mob mentality’. The more people there is, the dumber they become. Everyone’s individual ideas are squashed into one big idea that everyone consents to, meaning the end result is a mainstream idea.
This all changed with a book by James Surowiecki called ‘The Wisdom Of Crowds’. In it he revolutionised the idea of crowds and showed that they could in fact be smart. A bunch of individuals could in fact produce new, creative and innovative ideas, if the ideas were not ‘averaged out’, but instead nurtured.
...To Collective Mayhem
But this new train of thought was widely adopted and the collective intelligence trend boomed.
“I think a lot of people kind of went too far in the other direction, in saying, 'Wow, it's wonderful if we just let everyone participate on the web like in Wikipedia, everything will turn out great.', says Professor Thomas Malone, the Director of the Centre for Collective Intelligence at the Massachusetts Institute of Technology.
This saw the rise of a plethora of websites jumping in, eager to harness the power of collective intelligence, but not knowing how to go about it successfully. This meant many, many failures.
Trying to be all Organic
You can’t engineer collective intelligence by setting up an artificial situation, yet neither can you let the process be completely organic, and arise all on its own accord.
'There is the misconception that you can sprinkle crowd wisdom on something and things will turn out for the best. That is not true, it's not magic.' (Thomas Malone).
The answer, is that a middle ground approach is needed, as boring as that may be. The most successful types of collective intelligence require planning, rules, organisation and control first and then need to be let go and allow participants to contribute freely. This is a far cry from the sort of ‘melting pot’ mentality, where everyone’s ideas are thrown together and simmer automatically to generate an end product.
Too many Indians, not enough Chiefs
Then it is fair to say for quality collective intelligence to arise, some form of leadership is necessary. No matter how many people you have, nothing will get done if no-one knows what needs to be done. Some direction needs to be provided, otherwise all you gain is a bunch of loosely formed and competing ideas...a big indecipherable mess. James Gosling, the creator of software program ‘Java’ says that it is no coincidence that thousands of companies which use community knowledge are flailing and ineffective, while the company Apple is run by a “dictator with good taste” leads the way in technology design.
Use Me, Don’t Abuse Me
The web-based business ‘Threadless’ is a good example of how to use and not abuse collective intelligence. ‘Threadless’ gets users to submit their own artistic designs, which are then voted upon by other users and the top-rated designs are made into T-shirts which are then sold on the website. Collective Intelligence is about the community on Threadless, voting and discussing t-shirts designed by individuals. Dumbness of Crowds would be expecting the Threadless community to actually design the t-shirts together as a group. The following picture illustrates the 2 common results which occur when groups are asked to design something together, in this case the perfect dog:
Our Experiment
Our ‘Facebook as a way of generating ideas’ experiment fell victim to these common problems:
1. We got too excited too soon. We bought into the myth of collective intelligence as some kind of utopia where we would get a creative result with minimal effort by just ‘putting something out there’.
2. We created an environment that was in some ways too artificial (using a Facebook set up), but in other ways too organic (having little control over it). We modified and modified, but didn’t achieve the perfect balance which is needed to generate quality ideas!
3. We had no real leadership or direction to the collective intelligence.
4. This meant that the results gained were both types of ‘dogs’ from the above image. We got alot of average, mainstream intelligence in the form of viral videos. Whilst amusing, they didn’t generate any new ideas. If we attempted to put these ideas together we would have gotten the Frankenstein Dog - a mismatched collection of ideas with no glue.
For collective intelligence to work it cannot be left to its own devices, it needs work.
Experiment Conclusions
From the responses that we received, it was possible to draw the following conclusions:
- People will often respond much more readily to a question that relates directly to them, rather than a general question.
- People have a strong engagement with online content, and sharing this with others has become a natural activity that users enjoy doing.
- It is much more effective to ask an open question that allows anybody to respond, rather than try to force a response.
- Offering some form of payment is a good way to motivate people to respond.
Although at times we were successful in prompting a response, we failed to establish a collaborative environment in which a collective intelligence could be built. We had hoped that our participants would engage with each other and interact in a natural way as is common on Facebook. Unfortunately the way in which we set up our experiments created an artificial situation and did not facilitate the natural flow of ideas.
There are several ways in which we could improve our methods for future studies:
- We could send out an event invitation to all of our Facebook connections. This could outline the purpose of the study and what we hope to achieve. Doing this would mean that people who were interested would essentially volunteer to particpate. They would understand what was required of them, and interact with others that were involved with the study.
- We could use our statuses over a long period of time to ask a variety of questions. The short experiment that I did with this seemed to elicit a more natural response. We could even experiment with tagging specific friends in the questions to prompt them to respond.
- We could post our questions on our friend's walls. This would make them personally directed and hopefully the recipient would feel obligated to respond. One problem with this is that it is unlikely that others would get involved in the conversation.
- We could send questions out to individuals, asking them to add something and pass it on to some friends, and return it to us. We would lose control over who was receiving the questions, and hopefully receive some interesting responses.
Thursday, September 16, 2010
Time to Experiment...
Can Facebook really be used as a platform to collaborate with others and generate creative ideas?
From my Facebook network I chose a group of 30 friends, some who were already mutual friends and some who were not. This was with the idea of trying to have a variety of responses (with people not already knowing one another), but also hoping the pre-existing relationships in the group between friends would mean that they would build on and collaborate with each other’s responses.
I asked this Facebook network the following question:
What is the best idea you can think of for a movie?
There were only two responses for this question, and no discussion amongst participants at all. It was a disappointing failure. I realised this question was too general and quite frankly my participants could not be bothered to respond.
I then came up with a second question which I hoped would be more specific.
If you could choose any cast members for your own movie, who would you choose?
This question was more successful than the first. The responses I gathered were a range of celebrities and movie stars. There was humorous discussion among participants, and responses were a mix of short paragraphs of text and occasional image links. But I couldn’t help thinking that no real intelligence was being created in this forum.
Using the same group of participants, I then modified my question to: “Excluding movie stars, who would you choose to star in your film and why?”
This provoked an interesting increase in responses and a different type of response. One respondent posted this video to the Facebook thread, with the comment: “this kid as lead male, and his bro can be director”.
This first video sparked the uploading of more videos, such as this one with the comment “made by the roommate of some dude tripping out, best character I’ve ever seen, I’d watch him for 2 hours plus”.
Another response was this video with the comment "she's my pick for head writer/producer for my movie, soooo cute".
It was interesting that because the question was raised in an online forum, Facebook, my respondents were more inclined to provide answers that were understood by an online community - evident in the flurry of ‘Youtube celebrities’ in the viral videos that were added to the thread. It was also interesting to see how it only took one viral video for many more of my participants to follow suit and produce a similar response. It was almost like a ‘peer pressure’, or ‘fitting in with the crowd’ situation.
Several more videos were uploaded and discussed socially amongst participants. Most of the discussion was along the lines of “that was heaps funny”. Participants expanded on ideas, such as sending links to other videos starring the ‘World of Warcraft Kid’. The thread had approximately 5 days of high activity between the 2nd and 7th of September. In this period there were a total of 34 responses, many of these were just comments on other participants’ posts though. After this, activity slowed immensely and eventually died off until nothing more was added to the thread as of the 11th of September.
From my Facebook network I chose a group of 30 friends, some who were already mutual friends and some who were not. This was with the idea of trying to have a variety of responses (with people not already knowing one another), but also hoping the pre-existing relationships in the group between friends would mean that they would build on and collaborate with each other’s responses.
I asked this Facebook network the following question:
What is the best idea you can think of for a movie?
There were only two responses for this question, and no discussion amongst participants at all. It was a disappointing failure. I realised this question was too general and quite frankly my participants could not be bothered to respond.
I then came up with a second question which I hoped would be more specific.
If you could choose any cast members for your own movie, who would you choose?
This question was more successful than the first. The responses I gathered were a range of celebrities and movie stars. There was humorous discussion among participants, and responses were a mix of short paragraphs of text and occasional image links. But I couldn’t help thinking that no real intelligence was being created in this forum.
Using the same group of participants, I then modified my question to: “Excluding movie stars, who would you choose to star in your film and why?”
This provoked an interesting increase in responses and a different type of response. One respondent posted this video to the Facebook thread, with the comment: “this kid as lead male, and his bro can be director”.
This first video sparked the uploading of more videos, such as this one with the comment “made by the roommate of some dude tripping out, best character I’ve ever seen, I’d watch him for 2 hours plus”.
Another response was this video with the comment "she's my pick for head writer/producer for my movie, soooo cute".
It was interesting that because the question was raised in an online forum, Facebook, my respondents were more inclined to provide answers that were understood by an online community - evident in the flurry of ‘Youtube celebrities’ in the viral videos that were added to the thread. It was also interesting to see how it only took one viral video for many more of my participants to follow suit and produce a similar response. It was almost like a ‘peer pressure’, or ‘fitting in with the crowd’ situation.
Several more videos were uploaded and discussed socially amongst participants. Most of the discussion was along the lines of “that was heaps funny”. Participants expanded on ideas, such as sending links to other videos starring the ‘World of Warcraft Kid’. The thread had approximately 5 days of high activity between the 2nd and 7th of September. In this period there were a total of 34 responses, many of these were just comments on other participants’ posts though. After this, activity slowed immensely and eventually died off until nothing more was added to the thread as of the 11th of September.
Wednesday, September 15, 2010
Melissa - Experiment Four
As I was not getting much of a response from sending out questions via email, I decided to try some questions via status updates. This would allow anybody on my Facebook to respond if they had an answer that they would like to share, and it would not be a forced response. I created these statuses at times of relatively high activity.
The first status did not get an immediate response:
The first status did not get an immediate response:
I allowed an hour for responses to that question, before posting the next one. The second question received an immediate response from several people:
The following are the links to the sites that were suggested:
And the videos that were recommended:
The fairly extensive response to this question showed that many people have a strong connection with some form of online content, and that sharing this content is quite an easy thing. The responses were quite enthusiastic, and some participants responded more than once.
I also received several responses to my third status:
The following day my first status has received a response:
Although the responses provided may not necessarily be useful for creating video content, this experiment demonstrated that it is much easier to receive a response when the question is open to all Facebook connections. As the question was not directed to any particular individuals, only those that felt they had something to contribute responded. This also meant that most of the responses were serious answers, as people were not making jokes in order to feel as though they had provided some kind of answer.
Melissa- Experiment Three
For my third experiment I decided to ask a question that could potentially reveal the sort of content that would generate interest. I expanded the participant group again, and sorted them into groups that knew each other, to promote interaction. The question I asked was:
"What is the funniest thing you have ever seen?"
The response to this was varied, but most that did answer chose to send videos or links. One participant posted a picture:
"What is the funniest thing you have ever seen?"
The response to this was varied, but most that did answer chose to send videos or links. One participant posted a picture:
Two participants posted Youtube videos:
Two participants posted videos of their friends, that were hosted on Facebook.
Although the participants did not respond to the material others had posted, a social discussion thread did emerge. As this sent notifications to all recipients of the email, responses continued to be posted for several days after the message was initially sent.
Melissa- Experiment Two
For my second experiment, I decided to try to ask a question that was more personal to the respondents. I narrowed down the groups, selecting the participants that had responded to the previous question, and some others that I felt would be likely to respond. The question I asked was:
Two participants responded with single-word answers, which were "travelling" , and "tequila". I also received one video response to this question:
"What is the best idea you have ever had"
Interestingly, the respondents were more inclined to provide a written response to this, rather than linking to online content. The following is one such response:
Two participants responded with single-word answers, which were "travelling" , and "tequila". I also received one video response to this question:
In this case, the participants again did not interact with each other. Those that did respond did not add anything further to the thread.
Subscribe to:
Posts (Atom)