The other side of crowdsourcing: disadvantages, costs and complications.

In the last post, I analyzed the benefits for a company because the use of crowdsourcing. However, to have a complete view of crowdsourcing, it is necessary to talk about those costs, disadvantages and complications that its use involves.

In first place, the realization of a crowdsourcing initiative has associated costs:

1. Crowdsourcing involves both time and energy costs on platforms that allow the organization to effectively leverage the information given by the crowd. Needless to say that in cases where the crowdsourcer doesn’t use an existing platform, like Amazon Mechanical Turk or InnoCentive, the creation of one (or designing a website from where the crowd can participate) will result in higher costs.
2. In those initiatives that require the involvement of an online-community, a significant time cost that can delay significantly the release of any crowdsourcing initiative may appear. If the community does not already exist, there will, with time, dedication and care, to create and maintain it (one of the most difficult taks in crowdsourcing initiatives).

About the aportations of the crowd, 4 clear problems can be identified:

1. Sturgeon’s Law is fully present. This law, derived from the Pareto principle, indicates that 90% of everything is crud. So when the crowd gets involved in crowdsourcing initiatives, many of their aportations will represent information of little or no value or relevance. Obviously, all information must be filtered, which means a significant time cost.
2. Problems related to intellectual property (IP). Although it depends on each initiative, it is normal that the crowdsourcer retains the IP (Fiat Mio, project in which Creative Commons licensed was used is an exception). However, in the case of initiatives in which many users provide many solutions, a logo design contest for example, who will own the rejected designs? This problem can be solved explaining this situation clearly in the Terms and Conditions of the initiative.
3. There is no guarantee that the resulting end product (whether it’s performing a task or solving a problem) will be of sufficient quality or effective enough.
4. Related to the quality of the crowd contributions, a fact that sometimes occurs is that users try to cheat. For example, if a task is the translation of a paragraph from Castilian into English, what some people could do is to use online translators and present the obtained translation as their own (translation that usually is of low quality).

According to some authors, there are some ethical problems associated with the use of crowdsourcing:

1. Wexler (2011) states that those who promote crowdsourcing initiatives often try to distance themselves from the experts and professionals who are considered static people, against change, and which involve a significant cost. The problem here is not to get away from the experts getting close to the crowd, looking for innovation, but that companies use the crowd in a strategic way to distance themselves from those professionals who seek to profit from them.
2. Silberman et al. (2010) claim that crowdworkers are not an API (Application Programming Interface), but all too often are treated this way. Other authors are more blunt, like Postigo (2003), who argues that the working model that uses crowdsourcing (without quoting it explicitly because in 2003 the term did not exist), represents an emerging form of labor exploitation through the Internet.
3. In the case of crowdsourcing initiatives in which only one of the individuals in the crowd gets the reward, there is another ethical issue. This is the fact that by rewarding the work of a single individual, the work of the rest of the crowd goes unrewarded (though they knew that this could happend when they decided to participate) and in most cases is disposed of.

Regarding the rewards, in the case of initiatives in which all receive some financial reward (as in the proposed microtasks in Amazon Mechanical Turk), the crowdworkers  consider these rewards very poor (0.04$ for tagging 5 images, for example ), leading to unrest and crowd motivation decline.

Finally, there is the problem associated with the nature of the information provided by the crowdsourced. Whenever this information is sensitive or strategic in nature and do not want to made it ​​public, crowdsourcing is not the best way to work.

In conclusion, I consider essential to know which tasks can be performed using crowdsourcing and which not. Crowdsourcing is not the solution for everything.

References:

• Estellés, E., González, F. (2011) Crowdsourcing desde el punto de vista de la empresa: ventajas y desventajas de su aplicación en la resolución de problemas, III Congreso Iberoamericano SOCOTE y VIII Congreso SOCOTE “Tu + TIC = Innovación + Competitividad + Sostenibilidad” Universidad Politécnica de Valencia, 11-12 Noviembre de 2011
• Postigo, H. (2003) ‘From Pongto Planet Quake: Post-Industrial Transitions from Leisure to Work’, Information, Communication and Society. 6(4): 593–607
• Silberman, M. S., Irani, L., Ross, J. (2010) Ethics and tactics of professional crowdwork, XRDS, Vol. 17, pp. 39-43
• Wexler, M. N. (2011) “Reconfiguring the sociology of the crowd: exploring crowdsourcing”, International Journal of Sociology and Social Policy, Vol. 31 Iss: 1/2, pp.6 – 20