Crowdsourcing is evolving from exotic to mainstream as it becomes increasingly incorporated into the operations of major information companies, such as Thomson Reuters, Bloomberg, and others, to build and enhance their content collections. In the first phase of crowdsourced labor, services like Amazon’s Mechanical Turk sprang up as marketplaces for relatively low-skilled labor to do relatively simple tasks at pennies per task. These services demonstrated the potential value of crowdsourced labor, but lacked mechanisms to make it feasible on an industrial scale.
Now a new generation of service providers is providing tools that enable large companies to more easily take advantage of crowd labor within the context of their content operations. One of these players, WorkFusion, a company founded out of MIT research, illustrates the larger potential for crowdsourcing.
WorkFusion provides a platform in which crowdsourced labor is part of a larger toolkit enabling companies to manage internal as well as external sources of talent globally (e.g., Amazon’s Mechanical Turk, Elance/oDesk, and uSamp, among many others). Using the WorkFusion platform, companies can define a project by specifying each step, and then take advantage of both crowd labor and tools from WorkFusion’s library to automate some manual tasks. One of the potential powers of WorkFusion is that it can apply statistical techniques to the work executed by crowd labor to analyze which workers are best suited to certain types of tasks. Also, It can use its scale of pattern data to analyze work steps and rapidly create applications to replace some human labor with automation. WorkFusion’s approach has attracted customers from financial services, ecommerce, healthcare, and consumer packaged goods.
This more evolved form of crowdsourcing holds the promise of giving information companies alternatives to their own internal offshore operations or outsourcing firms. Both types of offshore operations often lack flexibility to handle peak loads and/or new types of work. Furthermore as labor costs change in the developing world, companies have needed to seek new outsourcing relationships in new countries with cheaper labor. New crowdsourcing platforms such as WorkFusion’s could allow information companies to take advantage of crowd labor on a flexible basis, to accommodate peaks in their work levels and/or take on short-term projects.
As crowdsourcing becomes both cheaper and more reliable, its real value will be not just in cutting the cost of existing content operations, but also in enabling vendors to improve their offerings in at least three ways. First, information companies will be able to collect and organize content that was previously too difficult or expensive to collect. Second, this more advanced crowdsourcing will make it possible to enhance content, such as through tagging, to an extent that was previously economically unfeasible. Third and perhaps most powerfully, crowdsourcing will make it easier for information companies to link information in separate databases, thereby driving entirely new applications. Many information vendors today face fundamental challenges when trying to uniquely identify and normalize entities from multiple sources. (Is the company called “General Motors” the same as the company identified as “GM?” Is the company called “IDC” in one source the same as the “IDC” from another source?)
Today, crowdsourcing tends to be restricted to simple, repetitive tasks. But more advanced crowdsourcing from companies like WorkFusion shows the potential for it to play a larger, more valuable role.