Milestone 3 taskforce poweridea1 qualitycontrolmanagers

From crowdresearch
Revision as of 16:58, 18 March 2015 by Cristinasarasua (Talk | contribs) (Created page with "'''Quality control for managers of complex, under-defined tasks ''' We propose an idea of quality control for managers of complex and ill-defined tasks. As it was exhibited i...")

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Quality control for managers of complex, under-defined tasks

We propose an idea of quality control for managers of complex and ill-defined tasks. As it was exhibited in "Expert Crowdsourcing with Flash Teams" paper, solving complex tasks in crowdsourcing environment require distributed leadership where requester might cooperate directly with the responsible for every part of a complex task. Following this insight, we address the challenge of picking the right manager by proposing a collaborative management of the tasks on the initial parts of the project. Just as in the enterprises, where some unsuccessful appointments of managers inevitably end up in failure, some of the appointed managers of crowdsourced projects will underperform and fail in theirs work. Moreover, given the distributed and loose control nature of crowdsourced projects, late detection of such managers will lead to increased project costs, and even more important, will cause substantial delay in a project schedule.

On the bright side, an average cost of hiring equivalent level manager is expected to be lower due to saved expenses. Therefore, we suggest a process where management decisions are done collaboratively by using the majority vote. In this way, an initial pool of potential managers will cross-validate the decisions of each others and will make it possible to recognize unsuitable managers in the first stages of the project. Moreover, as the tasks in complex projects are frequently interrelated, implementing this approach will explore potentially good managers in other, unexpected tasks. That will allow a requester to promote manager to another parallel tasks with subsequent promotion to a higher level managerial position. Suchwise, we pave the way for creating a working hierarchy equivalent to the one existing in the corporate world, with the difference that the detection of not competent managers in crowdsourced project is faster and cheaper. We believe that it is possible to detect low-quality managers on early stage of a project and maintain the quality by introducing periodic cross-validation process among managers.

From the power point of view, the decisions are made collaboratively, and the manager promotion is done in a fair manner. The intrigues and other irrelevant considerations stay outside the fence and promotion is done in a blind manner, based on the performance. Having said that, the presented approach raises a potential for tricking the process by collusion and coalition creation. We have ideas how to prevent this behaviour. However, at this point, we will leave this part out of the scope of this milestone.

A strong reputation system with strict rules can enforce both power and trust. By strict rules I mean: There should be a consensus among a set of individuals in order to grant or remove a “badge” of trust/skill. Also, we shouldn’t go from 100% and down, use earned badges instead.