The topic for this week was Crowdsourcing. Sincerely, I had never heard this word before reading articles, so I was really interested in understanding what the word means and its features.. But first, what does crowdsourcing means? Howe stated that crowdsourcing is the act of outsourcing a task to a “crowd”, rather than to a designated “agent”, such as a contractor, in the form of an open call.
The first article, by A. Afuah and C. L. Tucci, attempt to explain when crowdsource a problem-solving process could bring to better results than solve it internally or designing an external agent. If the problem-holder has enough knowledge to solve the problem internally, he is conducting a local search, if not, he needs to acquire that knowledge which could be found outside. In this case he should conduct an external search that sometimes could be time-consuming and expensive. The best advantage of crowdsourcing is that external agents self-select and try to solve independently a problem transforming a distant search in a local search. They found five key factors that influence the choice of crowdsource a problem or not.
- Characteristics of the problem;
- Characteristics of knowledge required for the solution;
- Characteristics of the crowd;
- Characteristics of Solution to be evaluated and evaluators;
- Characteristics of IT.
For each factor authors made some proposition. For example:
Proposition 1a: The easier it is to delineate and transmit a focal agent’s problem, the higher the probability the agent will crowd source the problem.
The second reading, by Lars B. Jeppensen and Karim R. Lakhani (2010), is a statistical study on the impact of marginality on problem-solving process in a broadcast search. In fact “marginal” people have no pre-existing knowledge or theory. Thus, they can solve a problem in different and novel ways. Two kind of marginality are individuated:
- Technical marginality: people are distant from the field of knowledge required to solve a problem;
- Social marginality: people are excluded from the field of the problem.
To evaluate the impact of marginality on the success of problem solving, they used two different regression models. Given the positive relationship between technical marginality and problem solving success, they found that successful solvers could bring new perspectives and heuristics into problem solving process. Furthermore, they studied women’s social marginality. In fact, they found that women are more cautious than men in submitting a solution, and that they do it only if they are pretty sure to have a good solution which could be the winner. Thus, they state that successful solutions are positive correlated with being women.
The third article, by T. W. Malone, R. Laubacher and C. Dellarocas (2010), defines the elementary parts, called “genes”, of collective intelligence systems. These genes can be combined and recombined to explore new ways of collective intelligence. They found that any activity needs some “genes” to answer to four key questions:
- What is being done?
- Who is doing it?
- Why are they doing it?
- How is it being done?
They found sixteen genes. For example: “How genes” for a create task (the actors in the system generate something new) are: Collection and Collaboration. The first one allows everyone to create content independently. The second one occurs when there are strong interdependences among single contributions that should be managed to perform a single task.
The fourth reading, by G. P. Pisano and R. Verganti (2008), analyses different forms of collaboration. They state that the kind of collaboration depends on two dimensions: Participation and Governance. The first one can be open or closed depending on whether or not everyone could participate. An open platform best suits if problems or situations could be separated into small “pieces”. The second one can be Hierarchic or Flat: Hierarchic if only one group has the last decision on a matter, while Flat when everyone can participate to decisions.
The two examples I used are: Yahoo! Answer and OpenStreetMap.
I think that everyone of you knows so well Yahoo! Answer. Basically, it is a database where every registered user can post its own questions or answer to other members’ questions. It’s the classical example of a collection gene: everyone is encouraged to participate or to find his or her question on the database. It’s easy, funny, easy accessible (it’s available in 12 languages) and well integrated with other Yahoo!’s services. But remember: the community is not made by expert. Furthermore, it’s not unusual to receive silly or joking answers.
OpenStreetMap is also known as Wikipedia of road maps. Basically, this is a crowdsource project born on 2004 which goal is to create a free and crowdsourced map of the entire world. It allows to registered users to contribute adding or editing maps. In fact, users can use their GPS peripherals, such as smartphones or laptop, to add maps, or they can use their “local knowledge” to add just details or point of interests. Its greatest advantage is that is distributed under ODbL license allowing users to use, share or edit freely maps. For this reason, many affirmed firms, such as Apple, Flickr, Joomla and Foursquare, employed it. Given its wide community that is now reaching 700000 users around the world, one more advantage is that it provides a constantly updated service. If you are planning to develop a map-based service this could be an optimal solution. Finally, this is an example of collaboration gene because everyone contributes independently, but there are strong interdependences among contributions that must be managed.