Crowdsourcing 101
Guest Post. Geng Tan, joined Mutopo forJanuary 2009, as part of MIT’s externship program. Geng Tan is a junior at MIT majoring in Mechanical Engineering and Management Science. Although he is a mechanical engineer by training, he is also interested in marketing and business development, especially through web 2.0 tools. In the past, he has interned at Hakuhodo Inc. one of Japans leading marketing consultancies and is currently leading up a crowdsourcing project at MIT related to education and technology.
Crowdsourcing
From the start, people have used the Internet to collaborate – the first research communities would use tools like e-mail to share information more quickly. In recent years, as global online participation has surged and connectivity costs have dropped, new forms of group collaboration have emerged as organizations try to harness the power of many connected people.
Earlier forms of these include the Linux “free-software” movement, but in the recent years more and more value, in the forms of Wikipedia articles, Youtube videos, Yelp reviews, Istockphotos’ pictures, SETI’s massive calculation power, just to name a few, are generated over the web through contribution of the “crowd,” the increasingly active internet users.
In this post, I would like to propose a framework for “crowdsourcing,” a term describing, according to wikipedia, “the act of taking a task traditionally performed by an employee or contractor, and outsourcing it to an undefined, generally large group of people or community in the form of an open call.” According to Jeff Howe, the author of Crowdsourcing, the foundation of crowdsource is the concept of spare cycles, the spare time people naturally spend on hobbies and leisure, such as playing baseball with kids, taking pictures, web-surfing and blog posting. Crowdsourcing is achieved by harnessing people’s spare cycles to generate value.
Some Examples
Businesses are paying more attention to this phenomenon because it has been shown time and again that these projects can produce products of superior quality (such as Mathwork’s competitions, Wikipedia, myStarbucksidea, Muji.net) and generate contributions such as content , computing resources or donations of massive scale (such as Yelp’s review systems, Istockphotos, Dell ideastorm, Kiva, Skype’s model of data processing) even compared to the most successful corporations. It also provides a few other advantages, such as reducing expenses associated with marketing, distribution as well as customer support.
I’ll first introduce how crowd source is used to generate quality and/or quantity of work, build a framework for businesses that attempt to crowdsource, and then list a few examples of successful crowdsourcing projects. I’ll conclude by listing the dos and the don’ts in crowdsourcing.
In any crowd sourcing project, users gather around an issue/task/topic and each work on a small portion of the project. Even though each person is working on a very small portion, the web allows for a massive accumulation of work. The crowd often forms a self-perpetuating community as well, and they can generate a lot of ideas, build upon each other’s ideas, and self improve the inferior ideas to form a better one. This almost resembles web games, such as the World of Warcraft, where users competitively build superior characters. As a result, there exist a myriad of characters and competitiveness nature of the system forces the superior characters to surface.
One key point to be noted here is that if the purpose of the project depended on diversity of ideas, such as in the case of brainstorming or producing reviews, the project manager needs to limit the interactions among the users. Too much interaction tends to form a trend within the community and such crowd-mentality trend to weed out the less main-stream ideas (imagine a forum where people nominated presidential candidates. Everyone chat about candidates and people agree on obama as a good candidate. Nominations of other candidate from this group seems very unlikely). However, if the project really wanted to produce a single quality idea, the manager want to encourage as much interaction among the crowd to encourage the crowd to build upon each other’s ideas (encouraging everyone on a forum to debate about who’s a superior candidate in the election.).
A proposed framework for making crowdsourcing work for your organization
1.Identify a crowd, a goal, and a method of contribution
The crowd can be preexisting or created. Pre-existing community can be internally marketed, and the existing rules can be imported.
If a new community is to be created it needs to be organized and encouraged. The new rules need to be created, and culture need to be formed. Relationship management is the key
The goal or goals needs to be clear and practical
Method of contribution needs to be simple and easy to use. Good user interface is the key, because it makes it easy for people to contribute their free time and wont get in the way of the primary experience, such as contributing and communicating.
2. Motivate contribution, motivate peer review/revision, collaboration, and consolidate the community
Motivate contribution by giving people the recognition for their work. People want to be heard, be appreciated, be recognized. This can be done by creating special “power user” status, giving front page coverage, etc. linked to specific measures of success.
Motivate peer review by encouraging discussion among users. “everyone’s equal on the web.”
Show that contributions make a difference and are appreciated – by other participants and/or by the organization.
Consolidate community by encouraging personal interactions. Easier for localized websites, such as yelp and facebook. Sponsored/hosted events may be helpful. i.e. yelp (yelp’s night’s out) and istockphoto (istockalypses).
You can check community health by looking at metrics such as number of posts, repeat traffic versus new, what percentage of people are contributing versus browsing, etc.
Also remember that people participate in different ways. For example perhaps only 1% of users will actively generate content while the majority might simply observe.
3. Market to a larger crowd and repeat 2.
Market via word of mouth. i.e. other people’s blog posts, forum posts, reviews, etc.
Market via scandalous/controversial/disruptive content. Lots of coverage from existing media, lots of people talk about it, lots of buzz.
Rely on sudden burst of publicity – people discovering an active community are more likely to try it out. If there are no signs of life already, no amount of promotion will help to seed.
4. Measure how you are benefiting
Does the community feel exploited? Its important to constantly monitor whether or not the organization is doing its part for the community in return for their contributions. If value is not fairly exchanged, you are likely to fail.
Use generated ideas for commercial purposes.
Revenues from advertising associated with generated content.
Get donation like Wikipedia personal messages, appealing to people’s interest, who other people that you are getting donated.
Ask your community how you should benefit and how they would like to benefit, too!
Some examples of these ideas in action
Manga Helpers
1. Identify a crowd, a goal, and a method of contribution
i.e. manga lovers, translation of Japanese manga, a manga data base
2. motivate contribution, motivate peer review/revision, collaboration, and consolidate the community
i.e. personal messages, grant special status, public acknowledgement, individuals can contribute different translations, and formation of clan/ranking of clans, paid trips to anime convention, etc.
3. Market to more crowd and repeat 2.
i.e. blog reviews, word of mouth, media coverage
4. Financial return
Gain enough data for commercial use, sell the data
Gain enough traffic, earn revenue from advertisement
Be disruptive enough to the existing big companies to be bought out
Get donation like Wikipedia style
Mathworks Competition
http://www.mathworks.com/contest/furniture/about.html
1. Identify a crowd, a goal, and a method of contribution
i.e. Coders, Superior Algorithm, Competition Style,
2. motivate contribution, motivate peer review/revision, and loosely consolidate the community
i.e. publicity via niche distributors, such as university prof, professional publications, blogs, etc. Open up the winning code, and frequently comment/update the competition website, generate hype among competitors and their friends via granting them publicity
3. Market to more crowd and repeat 2.
i.e. blog reviews, word of mouth, media coverage
4. Financial return on the capital
Gain good enough data for commercial use, sell the data
Build brand with students
Identify potential talent
Wikipedia
1. Identify a crowd, a goal, and a method of contribution
i.e. Specialists, Encyclopedia, Write articles,
2. motivate contribution, motivate peer review/revision, and loosely consolidate the community
i.e. ask professionals to write few articles, give people tools for revision, encourage discussion between professionals
3. Market to more crowd and repeat 2.
i.e. blog reviews, word of mouth, media coverage
4. Financial return on the capital
Currently just accepting donations, although there is constant suggestions to monetize using advertising.
Failed Examples
CurrentTV – TV programs created from feedback by users for the users failed because it overestimated crowd’s capacity to generate large professional grade content. Hence few contributions and little to talk about.
Assignment Zero – crowd source journalism and publish the content created by the crowd. Failed to build a seed community
So what can we learn from these?
Dos
Encourage community building among users
Think of ways to serve and help the users. In turn, they will help you
Separate the tasks into small distinct bits
Have a clear objective for the users to perform
Guide the crowd along the way
Listen to the crowd
Donts
Take the crowd for granted, they are your equal
Go against the will of the crowd, they have power
Leave the crowd alone, they need your guidance
Ask crowd to do too much at once, they don’t have time
Complicate the process, they don’t have patience
Exploit the crowd, they are not stupid.
In conclusion, crowdsourcing has promising potential for content creation (quality and quantity) and marketing. It has also been shown that ventures based on crowdsourcing, the whim of the crowd, can be self-sustaining and profitable. It also saves money and time for the existing companies. Effectively used, crowdsourcing can generate more ideas and form a strong brand loyalty for the company.
[UPDATE- we have expanded this idea to create www.colaboratorie.org]
Posted in: Crowdsourcing | Tagged: Crowdsourcing, dell, externship, manga helpers, mathworks, MIT, muji, starbucks, wikipedia | 8 Comments
8 Comments to “Crowdsourcing 101”
Drupal is an interesting example of a meta tool designed for crowdsourcing. It is interesting to note the “payouts” that are usually present in these projects. 1. the users are passionate or enjoy the topic subject. obssessives help form the core. 2. others can easily get value from the contributors contributions. 3. there is usually a phsychic, currency of recognition or ackknowledgement from the inside the community and the potential for external members to understand publicly that they are using a “public good.” I think a great phrase is “the miracle of the commons”, my phrase instead of the tragedy of the commons and there are key requirements, shared objetive, nearly universal acknowledgement of what constitutes a positive contribution to the common.
Yeah. We are talking more about value exchanged in terms of the contributors. But we have focused much less on the “silent majority” who get the “public good”. Do they somehow value this more? I like the “miracle of the commons” idea – are you saying that there are shared values around what constitutes positives contributions?
Hey Shaun (Stefanie Milner’s sister here).
Check out demandstudios.com, it is part of Demand Media. The (extremely successful) model so far is completely based on crowdsourcing. Through what we have learned by creating a large content studio, we have produced over 500,000 SEO based articles and videos to feed our own internal web properties,which have grown to be the #34 ranked network in Comscore with nearly 90 million uniques/month. Not bad, eh? Check it out- let me know what you think.
Wow. Hello Stefanie’s sister. Thanks for stopping by. We continue to find fascinating examples of how organizations and individuals benefit from crowdsourcing approaches – I feel like crowdsourcing implies that the people participating are like a utility. So trying to find a better way to describe a relationships which I think tends to be more symbiotic.
Anyway, enough rambling. Demand Media sounds very interesting – I will reach out directly to learn more.
Crowdsourcing is a great way to access collective knowledge. Perhaps Crowdsourcing 201 is how we use prediction markets to stimulate participation and increase the degree of knowledge accuracy.
Blog entry for reference: http://www.chrissparshott.com/2009/01/web-strategy-and-prediction-markets/
I enjoyed the link, but at the end, the post seems to be incomplete – looks like you referenced a DOD project, but the post was truncated.
Prediction markets are fascinating. I wonder about how you improve on some existing versions of say, financial futures markets? Or look at trending data from large data sets, like Google trends.
Our current interest is more on the side of trying to understand how best to organize and lead organizations that benefit considerably from the participation of non-employees in core aspects of what they do – from product development to support. It seems that a lot of the value comes from a handful of people and so perhaps prediction markets would benefit from enabling people to quickly ID the good predictors and strip away other contributors as noise?
[...] Crowdsourcing 101 [...]
[...] Crowdsourcing 101 [...]