The so-called “wisdom of crowds” can be found everywhere from a crowd booing an umpire’s bad call at a baseball game to viewers texting their favorite performer on “American Idol.” Today, the size of the crowd is measured at Internet scale given the number of people on Facebook (800 million) and the number of Twitter followers for celebrities such as Ashton Kutcher (8.2 million). As the crowd at the second annual CloudConf 2011 last week in San Francisco shows, crowdsourcing can also be a business model.
A few hundred people gathered at the UCSF Mission Bay campus south of downtown San Francisco to get information on the nascent, but emerging crowdsourcing industry, learn best practices for developing a crowd business and for attendees to compare notes on how their business is doing.
“I think it’s a great opportunity to collaborate with other vendors that are trailblazing a new market,” said Rob Vandenberg, CEO of Lingotek, which delivers translation services for enterprise content management systems. A U.S. company looking to expand to countries like China, Russia or Saudi Arabia will want its Web content or other marketing materials translated into each country’s native language. But language translation software, such as Google Translate, may not be enough to translate English accurately. In other cases it’s better for someone in the native country to double check the translation done by someone in the U.S., Vandenberg said. Often, some colloquialisms in the U.S. -- he gave the example of the expression “like a bat out of hell” -- don’t translate into other languages.
No one understands the pitfalls of inexact translation better than former President Jimmy Carter. On a state visit to Poland when he was in office, Carter told the crowd greeting him at the Warsaw airport that “I have come to learn your opinions and understand your desires for the future." Unfortunately, his State Department interpreter told the crowd in Polish, “I desire the Poles carnally."
Lingotek uses crowdsourcing to avoid such embarrassments by using three levels of translation, Vandenberg said, starting with machine translation for a first-pass level of translation, followed by translation by an assembled community of translators from the client company, such as employees, customers, resellers and other partners. As a final check of quality assurance, the content could be reviewed by professional translators who do this kind of work for companies all the time.
“Translation is just a great fit [for crowdsourcing], being inherently global in demand and global in participation by the actual contributors. It really is screaming for a Web-based delivery platform,” Vandenberg said.
Also at CloudConf 2011 was CrowdEngineering, a company whose business model is crowdsource customer service. Typically, customers who have a problem with their new smartphone, TV set or food processor call the company’s toll-free number and get routed to a distant call center. The crowd of people handling queries can be extended to include users of those products, said Gioacchino La Vecchia, CEO of CrowdEngineering.
“The killer application that we’ve found so far is customer service,” La Vecchia said. “You’re crowdsourcing with your customer base so that the customer base becomes a community and they become the workers.”
The advantage of this approach to the company is that the customers can handle 10 to 20 percent of the calls that would have gone to the call center, which saves the company money on its call center contract, he said.
Because crowdsourcing as a business is in its early stages, it’s hard to get a handle on how big it is or what various types of business models may emerge, said Carl Esposti, founder of Crowdsourcing.org. an fledgling crowdsourcing analyst organization.
Esposti identified four general crowdsourcing industry verticals: Ideation; expertise; task; and microtasks. Ideation is the process of generating ideas from a crowd, such as about what features or specifications should be included in a new product. Through each phase, or ideation, of product development, the product takes shape. The expertise model is like that of the customer service crowd where someone likely has an answer to each consumer question. The task model tries to solve a problem by assigning certain jobs to specific members of the crowd. And the microtask model solves one subset of the major task to contribute to the larger project.
“Crowdsourcing is a completely variable model,” Esposti explains “It’s suitable for ramping up very quickly and executing work because you can deploy the same amount of work to multiple workers who work in parallel. It allows for a lot more throughput in a much shorter period of time.”
The crowdsourcing organization is expected to complete a study within six to eight weeks further refining the types of crowdsourcing verticals around which the industry will be organized, he said. They’re relying on the wisdom of crowds to firm that up.
Comments