I recently found an interesting article on crowdsourcing and wanted to share with my fellow out-sourcerors. First, let’s look at what crowdsourcing is.

Crowdsourcing is exactly what it sounds like, outsourcing tasks to a large group of people. For those of you stuck underneath a rock, crowdsourcing has been utilized by large companies for years for such reasons as collaborating on product developments, understanding customer needs, creative solutions, or even task completion.

To help understand how crowdsourcing can be utilized, let’s look at Amazon.com’s website, http://www.mturk.com/, Mechanical Turk. This site allows subscribers to post tasks and outsource a job to anyone. It allows anyone to create an account, search for HITs (human intelligence tasks), accept a HIT, and complete the task. Once the requestor approves the work submitted the payment is deposited into an account for the worker. Such tasks include translation, website ratings, product ratings, and surveys. Current compensation on Amazon’s site ranges from a penny to approximately $13.00. Also vice versa, the site allows you to create HITs as well. It gives you a video tutorial and shows you how to optimize MTurk for the most effective results in asking workers to complete tasks.

Now that we are connoisseurs of crowdsourcing, let’s take a look at an article I found called “Crowd Sourcing the World” on TechnologyReview.com. The article explains the fundamental idea of crowdsourcing and Nathan Eagle’s newest project, txteagle.

Txteagle is outsourcing crowdsourcing, so to speak. This new project, similar to Mechanical Turk, will outsource and distribute tasks but via cell phones to the poorest and most underused work forces in the world. This will allow workers in foreign countries to participate in answering questions via text messages, audio clips, or video messages. Some of the benefits this project may offer are:

  1. Greater efficiency in the translation of smaller dialects and languages
  2. More reliable translations of smaller dialects will allow for localization of cell phones therefore making way for advanced technology
  3. Transcription/Translation would be cheaper then hiring an outside company
  4. The tasks provide an income to workers in developing nations
  5. More proficient ratings of local search results/relevancy allowing for market penetration
  6. Local blog/news and survey access for market research and up to date information

Although this may seem like the greatest idea since sliced bread let’s take a look at a few of the cons:

  1. Quality control is hard to, well...control. The level of accuracy for each answer will always be discretional since it’s a human answer. Even incorporating complicated algorithms to analyze the accuracy of an answer will never be 100% accurate.
  2. The targeted developing nations are regarded to as “developing” for a reason. Penetrating a third world workforce is great but not if no one can afford a cell phone or if they don’t have access to one.
  3. Also, most widely used cell phones in undeveloped nations would be very simplistic given the lack of technical infrastructure and knowledge. Would the capabilities of these third world cell phones be sufficient for a completion of such tasks? Or would the size of the screen or character limits inhibit the accuracy of the answers?

Although this new gig sounds fancy, I'm not completely sold that the main idea behind it is to help third world countries but more or less it's to keep costs down for companies.

So you tell me, is this another shrewd attempt at exploiting the poor by handing off elementary level tasks and providing little pay to workers? Or is this an inventive idea for a next generation look into unknown markets?

Let me know your thoughts!

Share To:

Jen Street

Post A Comment:

1 comments so far,Add yours

  1. While your thoughts about the limitations of Txteagle are accurate, there is massive opportunity for crowdsourcing to grow in developing countries in the future.

    Platforms such as 99designs, which ensures quality work by way of a contest structure, open up projects to anyone with access to a computer and internet access.

    Labor-as-a-service platforms like CloudCrowd are the next evolution of the Crowdsourcing 1.0 model. Their 2.0 features allow for more complex task structure and automatic peer review of work results.

    You're correct that algorithms alone can't ensure accuracy, but when you pair them with human review the results can be impressive.

    The infrastructure in these countries will have to improve at least somewhat before computer-based crowdsourcing starts to take off in developing nations. Crowdsourcing, however, may end up being the killer app that helps drive more local computer centers.

    Regardless, I don't believe that Txteagle is simply a shrewd attempt to exploit the poor.

    I believe it's an attempt to match work that typically couldn't otherwise be done, with a large group of people who can do it, and earn significantly more than they're earning today.

    ReplyDelete