Crowdsourcing: Harnessing the Wisdom of the Crowd Matthew Helms email@example.com
Copyright 2007-2008, Name of Student. All Rights Reserved. For copyright permissions, contact the author directly or Professor Amar Gupta (firstname.lastname@example.org) at The University of Arizona.
Crowdsourcing: Harnessing the Wisdom of the Crowd
Matthew Helms email@example.com Eller College of Management University of Arizona
Crowdsourcing, the act outsourcing a task to a largely undefined group, can be used to complete projects of various types. A model developed to assess crowdsourcing projects will be explored and then applied to several types of crowdsourcing projects to access either their likelihood for success or to look at ways to improve upon their current project organization or management. Issues and observations related to crowdsourcing will then be touched upon. Lastly, crowdsourcing as it relates to outsourcing topics, specifically the 24-Hour Knowledge Factory will be analyzed to assess crowdsourcing’s potential benefit to professional outsourcing.
The term “crowdsourcing” was coined by Jeff Howe in a Wired magazine article entitled “The Rise of Crowdsourcing” (Howe, 2006, “The Rise of Crowdsourcing”) which talks about the new trend in using the internet’s cheap pool of labor to create content, solve problems or tasks and conduct corporate R&D. Crowdsourcing involves breaking a task into subtasks that are then outsourced to an unknown public group of people. The basic concept is to outsource tasks that require human involvement to the general public. This is in contrast to outsourcing other tasks such as computing tasks being outsourced by grid computing projects like Seti@home (setiathome.berkeley.edu). Crowdsourcing borrows many of its features from the open source software development movement, but doesn’t necessarily apply only to technical projects. There are
Helms many historic and modern day examples of crowdsourcing projects outside of software development.
After describing several examples of different types of crowdsourcing projects, the paper will then use a model developed by Sami Viitamäki (Viitamäki, 2007) called FLIRT to analyze the successful design and execution of various crowdsourcing projects with the aim of understanding how to effectively use crowdsourcing to outsource tasks and activities traditionally done inhouse. Issues and observations related to crowdsourcing will then be touched upon. Lastly,
crowdsourcing as it relates to outsourcing topics, specifically the 24-Hour Knowledge Factory, will be analyzed to assess crowdsourcing’s potential benefit to professional outsourcing.
2. Crowdsourcing Projects 2.1 Longitude Prize
One of the oldest crowdsourcing projects was a project devised by the British government in 1714 who’s aim was to develop a device that determines the longitude of a ship at sea (Taylor, 1971). Navigation at sea without the aid of being able to see the shore is difficult and resulted in many lost ships and therefore lost resources and revenue for the British government. The project sponsors offered varying cash prize amounts, dependent on the accuracy of the device, to anyone who could develop or aid in the development of such a device. During the course of the project several inventors were recognized as contributing to a device to measure a ship’s longitude and therefore the project was a success overall. This type of open call requesting innovative ideas and content is at the heart of crowdsourcing projects.
The paper will now look at other specific examples of crowdsourcing projects in a modern day context applied to content creation, corporate research and development and outsourcing of physical labor.
iStockphoto is a website where amateur photographers can post their photos online and others can purchase the right to use the photo in their own work. The professional stock photography industry is comprised of professional photographers that take photos of a variety of objects and situations and then sell them to organizations that need photos for their own publications, but do not desire to hire a professional photographer to take custom photos. For example, if one wanted to create a website whose topic is learning and one wanted to have a photo of a tree on the site one could either hire a photographer to take some pictures of trees, purchase a photo from a professional photographer or one now has the option to purchase a photo from an amateur photographer that has posted their photo to iStockphoto. The price difference can be dramatic with the average cost of a professional photo at $150 and a similar photo from iStockphoto can be purchased for around $1. iStockphoto collects a percentage of all revenue generated as a revenue generation model (Howe, 2006, “The Rise of Crowdsourcing”). The popularity of iStockphoto has the potential of eliminating the professional stock photography industry.
2.2 CNN’s i-Report
CNN uses crowdsourcing in the i-report section of its website to generate content for their news network by allowing the crowd to upload video clips, photos or news stories of newsworthy events to their website; these video clips can also be used on CNN’s television broadcasts if the producers wish (http://www.cnn.com/exchange/). This use of crowdsourcing allows CNN to leverage the crowd to cover many more stories than would be possible if they only used their full time staff of in-house reporters. With the increasing number of people with cell phones that have photo and video capabilities, there is a virtual army of reporters that can be asked to capture and report on events around the globe on a continual basis.
InnoCentive is an online service that pairs corporations that need to solve scientific problems with individuals that have the ability to solve the problems. InnoCentive calls the corporations with problems to solve, “seekers” and the individuals that work on the problems and submit possible solutions, “solvers”. The solver that submits the best solution (as judged by the seeker) is paid from $10,000 to $100,000 for their innovative solution to the problem. The pharmaceutical firm Elli Lilly founded InnoCentive in 2001 and the site is now used by several Fortune 500 firms including Proctor and Gamble, Boeing and DuPont. The service helps mitigate the problem that current R&D costs are rising faster than sales revenue for certain industries (Howe, 2006, “The Rise of Crowdsourcing”). The outsourcing of certain, specific R&D tasks reduces the need to hire additional R&D staff and can leverage cross-domain knowledge by opening the pool of potential researchers to all academic and professional backgrounds. Based on a study by Karim Lakhani, researchers with expertise at the periphery of a domain are faster, on average, to find a solution than researchers in the domain (Lagace, 2006). Innovation happens at the intersection of disciplines and InnoCentive’s model successfully allows this intersection to occur.
Another key to InnoCentive’s success is its ability to work around two other problems within the existing R&D model: competition among researchers to publish and Intellectual Property (IP) concerns (Lagace, 2006). Even though science is based on openness and sharing of information, often times researchers are more concerned with publishing their research and this can get in the way of the sharing of ideas. Likewise, IP restrictions placed on researchers can hamper their ability to share ideas with colleagues and the scientific community.
InnoCentive addresses these concerns by posting the problem online for everyone to see and the solvers sell their rights to the idea when they submit it to the InnoCentive website. On the other hand, the solutions aren’t posted online and therefore a closed loop system is created. If the problem and solution were both posted online there would be the potential to iteratively improve upon the solution by incorporating feedback and further research done by others. Not posting the solution online is understandable given that the solution will be directly applicable in a corporate
Helms business environment. Lastly, another concern for corporations using crowdsourcing is the impact that this type of R&D can have on their corporate strategy. By posting research problems that they are working on online, they are revealing a lot about what they are working on and what their strategy is.
OnForce.com pairs IT service requests with IT service providers to complete IT related tasks. The concept is similar to the InnoCentive model except that the requested solutions are not necessarily innovative, but are service oriented. For example, someone in Tucson, AZ might need to install more memory in a computer, but doesn’t have the skills, time, etc. to complete the task themselves. Utilizing the OnForce.com website, they post their service request with the specifics of the task requirements as well as the payment amount and a registered service provider can respond and complete the task. In this way the IT task is outsourced to an undefined pool of labor allowing the requester to fulfill their work request in almost real time (median time of acceptance of tasks is 11 seconds as of November 14, 2007) and the service provider the opportunity to sell their services at market rates. Requesters utilizing the OnForce.com service have access to 11,000+ service providers nationwide with zero overhead which allows them to scale their operations as quickly as they desire (OnForce.com). Other similar sites are available for other types of tasks such as guru.com and elance.com which offer a wide range of services pairing requesters with service providers.
The FLIRT model (Viitamäki, 2007) will now be introduced in order to examine the likelihood of success of current projects based on their level of inclusion of the basic components of the FLIRT model. This evaluation process can be used in the future by project sponsors that are considering using crowdsourcing to complete some or all of the tasks of their projects.
3. FLIRT Model
In order to understand the different aspects of crowdsourcing and analyze why some projects are successful and some fail, the paper will use a framework developed by Sami Viitamäki, who is a marketing major at the Helsinki School of Economics and is writing his master’s thesis on the subject (Viitamäki, 2007). The elements that make up FLIRT are: Focus, Language, Incentives, Rules and Tools. These elements describe the major components that are necessary to harbor and support a successful crowdsourcing project.
Another major element of a successful crowdsourcing project is of course the participants. This paper will also take a look at Viitamäki’s breakdown of participant groups and how they interact to produce the final product. The groups identified by Viitamäki are Creators, Critics and Connectors and Crowds. An additional important component that needs to be present in order to ensure a successful crowdsourced project is the proper IP and legal groundwork in place and agreed upon by all stake holders before any content is generated or work done on the project. Figure 1 below is a graphical representation of the different elements of the FLIRT model.
Figure 1: The five elements of the FLIRT model and the IP and Legal framework and how they interact with participant groups (Adapted from figure by Viitamäki, 2007) 7
The focus of the project must be carefully planned and articulated to achieve success. The focus includes the strategic objectives and goals of the project. Viitamäki argues that there are three elements of the focus that are necessary with regard to collaboration in the project: area, scale and depth. The project’s objectives will dictate the intensity of these three areas such as what area of business will the project encompass, what ideal number of participates and what level of contribution you are expecting of them and what level of control you are willing to give them. If these boundary conditions are not established in advance then the chance of failure is greatly increased. Additionally, if the target participates are not identified and involved, then the project could also be unsuccessful.
The language element of the project is crucial in communicating with participants. Since what you are looking for is immediate contact with a targeted segment of the population, there is no room for generic messages. This is personal one on one communication and participants will see through any non-authentic messages or language. Communication should also be transparent and show that you understand and respect the community of participants. The language should mirror the tone and vocabulary of the participants and display a real level of caring and involvement. Another potential problem is which spoken language to use for the project (e.g. Italian, German, English, etc.) If the chosen language is English for example, the project might have an inherent bias towards English speaking participants and might not yield the desired result. Communication is crucial in crowdsourcing projects and the language used to communicate is an important part of a successful project.
There must be some type of incentive to compel participants to be involved in the project. Viitamäki outlines two main types of incentives, intrinsic and extrinsic. Intrinsic incentives include a challenge, a creative opportunity, to satisfy a curiosity, or for learning and fun. Extrinsic incentives can include both material and immaterial rewards. Immaterial rewards such as fame, recognition, access to resources or material rewards such as the opportunity to own beta products and try beta services, or monetary rewards. Adequate incentives for participation are required to ensure that the project has a critical mass of participants necessary for success.
Well defined rules and guidelines for the project must be defined before the project is initiated in order to guide the participants towards the goals and objectives of the project. Rules should be laid out for: initiation, interaction, intellectual exchange, manufacturing constraints or any other boundary conditions involving policy, scope or physical restrictions.
Rules for initiation focus on what can be done anonymously, what needs to be done by a registered participant and when such registration needs to occur. Interaction and intellectual exchange rules should be established depending on the desired outcome as well as the nature of the project. If the project is a competitive individual solution to a given problem, then interaction before the submission deadline might need to be limited. On the other hand, if innovation and creativity are the main objectives and there is no specific reward for being the best solution then a collaborative approach can be used that maximizes interaction. Communication protocols and guidelines also need to be established to ensure that there isn’t abuse of participants from other participants (e.g. verbal abuse or non-professional language). One issue not addressed in by the FLIRT model is the issue of intellectual property and who the owner will be of any creative products produced. Before launching the project, these rules, policies and guidelines need to be in place to avoid problems later in the project that might jeopardize the success of the project.
Once all the above issues are addressed, tools need to be in place for the project to proceed. Depending on the scope and objective of the project, some of the following tools will be needed: platform, creation tools, skills and knowledge and analysis tools.
The platform can either be custom built, purchased from a third party or a hybrid of the two. Most crowdsourcing projects leverage the crowd of web users and therefore the platform is usually web based. This type of platform needs not only a software based solution, but also the hardware and infrastructure to support the expected number of participants. If the project has a creative element then tools for encouraging and allowing for creativity need to be in place. These tools should be user friendly and have the capabilities to create the type of products that fit the goals of the project. If the project is very specific to a certain industry or topic, specialized skills or knowledge might be required to participate meaningfully in the project. Special attention should be given to this target participant group to ensure that you not only attract these participants to the project, but also that the tools you supply are familiar and similar to tools that they are already accustomed to. All of these tools are required to host a successful project and ensure that the full potential of the participants is reflected in the final product.
The FLIRT model developed by Viitamäki is a useful tool when analyzing crowdsourcing projects and the elements necessary for a successful project. Figure 2 below shows how the five elements of the FLIRT model and the legal and IP framework are strategic, tactical or technical in nature.
Figure 2: The five elements of the FLIRT model and the Legal and IP framework and how they apply to the strategic, tactical and technical areas of a crowdsourcing project (Adapted from figure by Viitamäki, 2007)
3.6 Participant Groups
Obviously in any crowdsourcing project, the crowd is a crucial element. Viitamäki outlines three main groups of participants that play distinctive rules (although some participants can play a role in more than one group during the project). These groups include creators, critics and connectors and crowds.
Creators play the role of generating original content for the project. They often times compete for the best solution and offer creative ideas to the project. They are involved in the project for the challenge, learning, fame, recognition and explicit rewards. To attract this group and keep them motivated, the project must provide an incentive that is applicable to their community or
Helms demographic and creative freedom and the right tools must be given for them to create the original content.
Critics and Connectors are involved in the community and the communication surrounding the community. They help to spread the word about the project and aim to influence others. Critics emphasize their opinions and want to communicate with large groups of people and are therefore attracted to the project if you offer this environment for them. Other items necessary to attract critics are the ability to interact with participants and a truly transparent and authentic channel of communication.
Crowds are the last element of crowdsourcing projects. They usually interact with the project at a low level, but have a huge influence on the success or failure of the project. It is best to use them in key strategic events where their input can be exploited effectively. These participants are often consumers of the product or service that the project is focused on or a similar product or service. Crowds might not communicate with other members of the project’s community, but they converse with their friends and are a powerful communication tool (either positively or negatively) because of this word of mouth form of information conveyance. The crowd ultimately decides what products and services are valuable and which are useless. In order to attract these strategic participants it’s important to make participation in the project very easy, show their influence in real-time and attempt to draw in their participation to deeper levels. Participation in these types of communities is on the rise and the goal of any crowdsourcing project is to maximize participation at all levels. In order to accomplish this goal, the project should be designed to eliminate barriers to participation and be designed to draw participants into deeper and deeper levels of participation.
If all of the elements of the FLIRT model are thoughtfully put in place and the right attention is placed on the three groups of participants, the likelihood for success will be greater than if the project designers don’t look explicitly at these issues and components of the project.
4. FLIRT Model Applied to Specific Cases
The focus of the paper will now turn to specific examples of crowdsourcing projects and how the FLIRT model can be applied to the project in order to either evaluate some of the potential reasons for its success or failure or gauge its projected success if the project is currently in its infancy.
4.1 Content Generation 4.1.1 iStockphoto.com
The success of iStockphoto when looked at using the FLIRT model is predictable. The focus of the project was clearly defined: to provide a platform for amateur photographers to post stock photos for others to purchase. This project is a content generation endeavor with a high degree of customer collaboration. The control of the project is mostly in the hands of the owners of the site and the participants are given specific tools for them to upload photos. The language of the site is appropriate for amateur photographers and simple enough to be enticing and not intimidating. The incentives are two-fold; in addition to the monetary reward involved with the participants’ photos are sold, there is the fame of being selected as the “artist of the week” or by being in the “Most popular” list. The site appeals to the large number of hobbyists that enjoy photography, but are not professional photographers and satisfies their need for a creative and fun outlet. iStockphoto has laid out the ground rules for what is appropriate and what is inappropriate material, terms of the license, IP restrictions, etc. and they have in place simple effective tools that allow for creativity and reduce the barriers to participation.
The groups of participants are also addressed with ways for all participants to be involved in the project. Creators are obviously a major portion of the projects participants, but the critics are also used to rate the photos and post to the community discussion board and are attracted by the ease of use of the website and the ability to communicate with a large audience through the site.
Helms Finally, the crowds are drawn in by the ability to browse photos and read the comments and posts of other participants.
4.1.2 CNN’s i-Report
Looking at CNN’s i-Report project through the framework of the FLIRT model, we see that most of the components are present, but some of the critical aspects might be missing. The focus of the project is clearly the collection and dissemination of news items and falls within the popular realm of crowdsourced content creation. The depth of participant control is minimal beyond the content creation and the scale of collaboration is also minimal.
The language used on the site is appropriate and CNN’s brand image and position in the journalism industry makes the component of language to convince participants of their interest in the project a moot point. The incentive for the participant is the fame of having their video, photo or story posted to CNN’s website or even aired on CNN’s television network. There is even the ability to upload a photo of one’s self that can be linked to one’s submission on the website. One of the biggest intrinsic values to the participant is the satisfaction of sharing the news, something that most people enjoy doing. According to their website, no material rewards are offered or given to participants.
The rules for participants are clearly laid out on the website and the terms and conditions must be agreed upon before participation is possible although it seems that the rules favor CNN more than the participant. The motivating factors for participant involvement are not clearly articulated anywhere on the website; this might limit the level of participant involvement.
The tools are in place for participants to upload their contributions and seem to be fairly user friendly. There isn’t much room for creativity and the platform restricts the participant’s contribution to only plain text (which presumably is edited and formatted by CNN employees) and file uploads of video and photos. More participant involvement might be garnered if CNN allowed
Helms participants to be more creative and take more ownership of their news items by providing a platform that allows them to format their stories or edit their video footage before uploading.
I-Report in the context of the FLIRT model has many of the components present, but could be better modeled to take advantage of the crowd. Based on CNN’s management of the i-Report project and CNN’s position in the journalism industry, it is not considered a critical project by CNN and will not be a determining factor in their success in the market. That said, crowdsourcing of news reporting is currently underutilized and has the potential to create a large shift in the industry and upset current market leaders such as CNN.
4.2 Corporate Research and Development 4.2.1 InnoCentive
Applying the FLIRT model to InnoCentive we can see that they have most of the components in place with the exception of perhaps some of the groups and a clear focus. The focus is obviously the business area of corporate R&D, but this might be too broad an area to attract the types of participants required to solve large scale corporate R&D problems. The language is in line with their target participants and shows them that the service offering is serious. The monetary rewards are significant when compared to other crowdsourcing projects, but so it the time requirement on the part of the participant. There are several intrinsic incentives involved in InnoCentive projects such as a challenge, learning opportunities and appealing to participant’s creativity. The rules are well defined and are inline with the goals and objectives of the project and the tools are in place for both the solvers and the seekers and no particularly innovative tools are required to post problems and solutions.
Where InnoCentive might fall short is in its ability to foster a community and provide adequate opportunities for participant involvement. The creators are obviously involved as solvers, but the critics, connectors and crowds are not involved in the project. This might create a problem in the future if InnoCentive wants to scale the idea and more participants are needed to act as solv-
Helms ers since the current model does not have levels of involvement to draw in participants to the deeper levels.
4.3 Crowdsourced Labor 4.3.1 OnForce.com
Looking at how the FLIRT model applies to OnForce.com we see that they have many of the components completed. The focus of their service is the well defined market of IT services and they list the service areas provided on their website: Computer Desktop Repair, Computer Laptop Repair, Computer Peripherals Upgrades, Consumer Electronics Installation, Network Support, POS Installation, Printer / Copier Repair, Security Upgrades, Server Support, Software Installation, VoIP / Telephony and Wiring & Cabling. The language used on the site is clear and easy to understand for both the requester and the service provider and uses the correct terminology and certification titles reinforcing its validity and reliability.
The extrinsic incentives for both parties are clear; the requester can efficiently and economically get its IT service request fulfilled and the service provider gets paid. The intrinsic incentives are a little more subtle. The value to the requester is that they don’t need to maintain a workforce of trained IT service providers and have instant access to a nationwide pool of labor. The value to the service provider is the flexibility of being able to choose when, where and for what price they would like to work. Based on the number of requests posted and the payment provided for each request it appears that a skilled service provider could scale their income from OnForce.com from part time to full time employment without much difficulty. This flexibility allows both the requester and the service provider great leeway in their involvement in the OnForce.com community.
The rules for involvement are well established and easily accessible for both the requester and service provider on the OnForce.com website. Many of the show stopper issues such as quality and reliability of service and payment issues are addressed with an FAQ page for both parties. The tools in place on the website are easily used to both post service requests and offer to fulfill 16
Helms the requests. The site is easily navigable and clearly laid out for both the requester and the service provider.
The groups of participants are adequately represented to sustain the success of OnForce.com with creators, critics and connectors and crowds being involved. The creators are the requesters and service providers and their involvement has been discussed above. Critics and connectors are able to interact with the community through either the OnForce.com blog or discussion group. They can comment on the blog posts or add to the discussion group by commenting on posts or posting new comments themselves. This level of involvement can easily become deeper if the discussion group post that they add to or in other words “solve” is accepted by the community. If they see their success in the discussion group, they might want to apply their knowledge and skills to actual paid service requests. On the other hand, the crowd can be drawn in if they browse the discussion group or blog and find useful information (or provide solutions as discussed above). If they see that the community of service providers is knowledgeable and see similar service requests to requests that they need fulfilled being posted to the website, they might decide to post a service request themselves.
By looking at OnForce.com in the framework of the FLIRT model, we can see that the creators of the website have many of the components of the model in place and therefore the service has a high probability of success. An aspect of the service that needs to be constantly monitored to ensure success is the quality and supply of service providers. If the quality of service can not be maintained, the word of mouth of the crowd will quickly degrade the site’s credibility. Lastly, if the service’s popularity outstrips its ability to attract more service providers then the median time to acceptance for new service requests will increase and some of the value to requesters will be lost.
5. Overview of Crowdsourcing Evaluations
Figure 3 below gives an overview of how the given examples of crowdsourcing compare to the model presented in this paper. We can see that some of the more successful projects have all the elements present whereas some other new projects are lacking or are deficient in some areas.
Figure 3: Overview of Crowdsourcing Evaluations
6. Other Crowdsourcing Issues and Observations
In addition to the components and concepts outlined in the FLIRT model above, there are other issues to consider when crowdsourcing a project. Many lessons can be learned from open source software development that can be directly applied to crowdsourcing projects.
To begin, the right management for the project must be in place before beginning the project execution. Without the right structure and management for the project, the groups of participants 18
Helms will quickly become disillusioned with the project and leave. Experts with experience in the area of the crowdsourced project are needed to manage and guide the project towards a successful conclusion. Additionally, the tasks that need to be accomplished for the successful completion of the project should be clearly defined and self contained if possible. These self contained tasks will help ease the management of the project and will require less supervision and therefore overhead for the project.
Lastly, some tasks don’t lend themselves to crowdsourcing such as when IP considerations outweigh the potential benefits of crowdsourcing. On the other hand, if innovation is a key component of the project, it should be opened up to as many participants as possible in order to leverage the potential of the crowd and increase the probability of capturing the innovative solutions that are found on the fringes of a domain of knowledge as was discussed in the InnoCentive example above.
As discussed in the FLIRT model, fostering a sense of community and having all the components of the platform in place for the participants is paramount. The crowd and their participation in the project is the largest asset in crowdsourcing projects and therefore all of the needs of the creators, critics and connectors, and crowds should be in place in order to ensure successful involvement of these participants. If any of the three participant groups is neglected, the sustainability of the project will be limited.
There are several aspects of crowdsourced labor that differ from traditional labor and these aspects much be carefully considered when approaching a crowdsourcing project (Howe, 2006, “5 Rules of the New Labor Pool”). The first difference is that the crowd is dispersed and therefore the tasks need to be able to be completed remotely unless the task its self is geographically constrained as in the case of OnForce.com. The second difference is that the crowd generally has a short attention span and therefore if tasks can be completed within 30 minutes the probability of someone from the crowd (versus a creator) completing the task is greatly increased. The third aspect of crowdsourcing participants is that the crowd is full of specialists. The internet is a very efficient aggregator of information and allows the specialists in the crowd to easily connect with other specialists or participate in a crowdsourcing project where their unique specialty is appli-
Helms cable. The fourth item to be considered is that the quality of output of the crowd varies greatly with only a small fraction of the output useful to the project. On the other hand, the crowd is also very skilled at filtering this output and pushing the quality work to the top by using rating systems such as that used by YouTube.com and other content creation crowdsourced projects or edited by users such as the Wikipedia.org project. All of the aspects of the crowdsourced labor pool must be considered when planning a crowdsourcing project to ensure success.
7. Outsourcing Issues Relating to Crowdsourcing 7.1 The 24-Hour Knowledge Factory
The concept of a 24-Hour Knowledge Factory as presented by Dr. Gupta (Gupta, 2007) is directly relevant to crowdsourcing in that crowdsourcing can leverage many of the same advantages that the 24-Hour Knowledge Factory employs by shifting work around the globe following the sun.
Crowdsourcing projects that are not geographically confined such as the case of news reporting outlined in the example of CNN’s i-Report lend themselves to using the model of the 24-Hour Knowledge Factory. News develops around the globe during all hours of the day and night and a reporter working on a story can only work so many hours before they need to sleep. By combining the model of the 24-Hour Knowledge Factory with that of crowdsourcing, geographically dispersed teams of reporters could monitor and report on a story around the globe, around the clock. These teams could be largely composed of participants from the crowd and managed by a smaller core group of professional reporters that would act as advisors and editors of the content produced by the crowd. Participants could either create a new story or add to existing stories based on their first hand experiences.
This would expand on the Associated Press’ model of employing dedicated reporters around the globe. Even though the Associated Press is a great benefit to many news outlets, it is still limited by employing only around 3,000 journalists and reporting the news in only four languages (http://www.ap.org/pages/about/about.html). A correctly constructed and managed crowd20
Helms sourced news reporting project that operates globally could provide better coverage and more content than any news agency today.
Other crowdsourcing projects such as Wikipedia.org could also benefit from pairing the 24-Hour Knowledge Factory with crowdsourcing. Wikipedia has many pages that cover current events or other topics that are evolving quickly. As the event unfolds, Wikipedia participants could update the page during their daylight hours and this updating progress will continue around the globe with the sun. When the participants check on the page when they wake, they will find that the page has been updated by other participants in other time zones. This type of crowdsourcing project that takes advantage of participants in different time zones can greatly enhance the advantages of crowdsourcing.
Another similarity of crowdsourcing and the 24-Hour Knowledge Factory is the flexibility afforded participants involved and both models’ ability to leverage specialized knowledge of participants. The 24-Hour Knowledge Factory eliminates the need to work during nighttime hours, allows workers to telecommute, encourages specialization of worker’s skills and allows professionals from different cultures and geographic locations to interact and engage in tasks (Gupta, 2007). Crowdsourcing has many of these same attributes and advantages in the addition to the ability to involve as many experts as the project management is able to attract from the crowd.
These common aspects of crowdsourcing and the 24-Hour Knowledge Factory highlight the potential applicability of crowdsourcing to professional outsourcing activities. New labor paradigms such as crowdsourcing of labor and services blur the distinction between amateurs and professionals. This distinction will continue to erode as workers continue to specialize, are able to telecommute from any geographic location, have increased flexibility in their working hours and shift towards more freelance work.
Crowdsourcing of projects has been successfully used to develop innovative ideas in the past and present crowdsourcing projects range in scope from corporate R&D to onsite technical support. The FLIRT model outlines the major components necessary for a successful crowdsourcing project. The focus of the project should be clear to participants, outline the scale of collaboration of participants and specify the depth of customer control over the project. The language used in the communication with project participants should convey the authenticity of the project’s managers and their knowledge of the business area as well as show the participants that the project sponsors are involved in the project and have a stake in the outcome. The incentives for involvement should include both intrinsic and extrinsic incentives such as a challenge to the participants and material or immaterial rewards sufficient enough to ensure the involvement of the desired participants. The rules of involvement should be clearly communicated and specified before the first participant involvement occurs. Lastly, the tools required for the project, such as the platform and any special creation or interaction tools, should be easy to use and familiar to the target participant group.
The groups of participants fall into three categories: creators, critics and connectors, and the crowd. Involvement of all three groups is necessary for a successful crowdsourcing project with the creators generating content and innovative ideas, the critics and connectors spreading the word about the project and the crowd deciding what is valuable and what is not. The platform should be designed in such a way as to not only involve all three parties, but also to draw them in to deeper levels of involvement such that members of the crowd can easily become creators.
Crowdsourcing can successfully be applied to many areas of outsourcing such as content creation, R&D, and onsite labor. In addition to outlined model, other resources are required for the successful completion of a crowdsourcing project such as adequate management of the project, experts to guide the crowd, clear task definition and careful consideration of IP and other legal issues. Lastly, more traditional outsourcing models such as the 24-Hour Knowledge Factory can be combined with crowdsourcing to leverage the advantages of both models.
“Associated Press”, www.ap.org Boutin, Paul. “Crowdsourcing: Consumers as Creators.” BusinessWeek July 13, 2006 Gupta, Amar, "Deriving Mutual Benefits from Offshore Outsourcing: The 24-Hour Knowledge Factory Scenario" (March 2007). Available at SSRN: http://ssrn.com/abstract=968761 Ho, Michael. “Assignment Zero First Take: Wiki Innovators Rethink Openness.” Wired May, 3 2007 Howe, Jeff. “5 Rules of the New Labor Pool” Wired June, 2006 Howe, Jeff. “Crowdsourcing: Tracking the rise of the amateur” Howe, Jeff. “Did Assignment Zero Fail? A Look Back, and Lessons Learned” Wired July 16, 2007 Howe, Jeff. “Gannett to Crowdsource News” Wired November 3, 2006 Howe, Jeff. “The Rise of Crowdsourcing.” Wired June, 2006 Lagace, Martha. “Open Source Science: A New Model for Innovation” Harvard Business School November 20, 2006 Lakhani, Karim R., Lars Bo Jeppesen, Peter A. Lohse, Jill A. Panetta. “The Value of Openness in Scientific Problem Solving” Harvard Business School October, 2006 McDonald, Mike. “Lost in the Crowd: How Crowdsourcing can Backfire on a Business.” RoboticBlue “Mechanical Turk”, Amazon Tapscott, Don, Anthony D. Williams. “Innovation in the Age of Mass Collaboration.” BusinessWeek February 1, 2007 23
Taylor, E.G.R., The Haven-finding Art: A History of Navigation from Odysseus to Captain Cook, Hollis & Carter, London 1971 Viitamäki, Sami. “The FLIRT Model of Crowdsourcing - The Updated Model and Background” May 6, 2007 Whitford, David. “Hired Guns on the Cheap.” Fortune Small Business Magazine Issue March 1, 2007 Worthen, Ben. “Crowds May Be Wise, But Are They Productive?” The Wall Street Journal July 16, 2007