I’m always interested in news ways of getting work done. Crowdsourcing is definitely new. It employs the power of the internet to find thousands (even millions) of workers to complete large but simple projects. Example are:
- sales force example: add twitter links and company names to a list of contact leads
- website example: removing explicit images from a website
- ecommerce example: improving product recommendations based on what users have bought in the past
All of these example are better done by human workers than by algorithms. They are large scale and repeatable with little no changes in the steps from one task to the next.
The two alternatives to crowdsourcing are insourcing (i.e. traditional hiring) and outsourcing. One big reason that crowdsourcing is better is because it’s very reactive to the amount of work that has to be done. With crowdsourcing we pay workers by the task, not by the hour. This means we only pay for workers when we need them. We can scale-up and scale-down very elastically to demand. Hiring full-time or contract workers takes 2-4 weeks to on-board, and outsource companies bind customers to minimums requirements that.
The graphs below describe this concept:
Crowdsourcing is not perfect. There are downsides too. Firstly, large scale crowdsourcing today is for simple and repetitive tasks only. No one has figured out how to guarantee quality for large scale projects with difficult tasks. An example of difficult tasks would be diagnosing a broken bone based on an x-ray. Perhaps this is a long way off. Secondly, crowdsourcing is a nascent industry and the supply of workers is not infinite. Even the best companies in the space are challenged to scale to customer demand.