Some analyses of crowdsourcing have described it as a form of “social” network. But is crowdsourcing really social?
Can paid work be social?
Government, nonprofit, and even some commercial firms use volunteers for crowdsourced tasks. Volunteer-based crowdsourcing is inherently more social than the semi-anonymous piecework of commercial crowdsourcing, because the workers clearly share a common interest or goal. For paid crowd workers, though, the possibility of social connection is limited. Unlike typical office workers who must interact because they’re in the same room, crowd workers are spread all over the globe and come from different regions, backgrounds, and cultural traditions. What do they have in common? Boredom? The need to supplement their income? Neither are compelling enough to foster bonding or “social” behavior among otherwise anonymous members of such a large, disparate group.
Can you be social if you are anonymous?
Crowdsourcing workers are typically anonymous to their employers and each other, the antithesis of true-identity social as mandated by Facebook and Google. And not for nothing–it’s difficult to socialize anonymously. In spite of their anonymity, though, “the crowds” do talk amongst themselves about projects they’ve worked on through message boards, sharing information that’s important to them about the good and bad “actors” that post crowdsourcing projects. These comments are the B2B equivalent of “liking,” “being a fan,” or otherwise recommending a firm. Crowdsourcing message boards are also important forums for sharing best practices for various types of crowdsourcing processes. So, the crowd is not inherently social, but it is almost certainly going to become more and more interconnected and social over time.
“Fans” as a “crowd?”
Marketers are increasingly taking groups of social “fans” and using them as crowds. “Internal crowds” can include all employees of a company, all customers of a company, all students at a school, or all brand fans on Facebook, to name just a few possibilities. Leveraging these disparate groups has gradually become a favorite sport among social media marketers. Ben & Jerry’s ice cream, Heineken beer, Lay’s potato chips, and Citroen cars have all successfully conducted new product efforts using their customers’ social participation. Perhaps one of the most publicized successes belongs to Mountain Dew, which solicited new product ideas from their internal fan “crowd.” After two million fans participated, the company rolled out the most popular of the three proposed new products to be developed by the crowd. More recently and perhaps more importantly, though, a Mountain Dew campaign proposing to get a new name for a drink from the same crowd was hijacked. The highly social fans played off each other’s joke names, enjoying the fun until the project was cancelled. Needless to say, corporate marketers at Mountain Dew were not amused.
This kind of mischievous behavior from a crowd of fans is embarrassing, but unlikely to cause much lasting damage—Mountain Dew drinkers still reach for their citrus-caffeine pick-me-up at about the same rate, much as McDonalds customers kept buying burgers during the quickly derailed #McDstories effort. For B2B operations, this kind social behavior in the crowds can lead to gaming the system and the loss substantial amounts of money paid out for bad data.
Does the “looping” approach to crowdsourced projects make them social?
The truest distinction regarding the social aspects of crowdsourcing comes when deciding what project is appropriate for specific workers. In the case of the collection of relatively simple data points, “looping” is employed. Looping is when a consensus response is required—in most cases, 2 out of 3 is enough warrant accepting the consensus response. In these cases, crowdsourcing is absolutely a social phenomenon. The very definition of looping means using more than one worker to seek out, enter, and ultimately use consensus values to validate responses. However, when data collection is more challenging or requires specific skills, such as categorization or moderation, a “Master” worker is employed. A master in crowdsourcing is the best-possible qualified worker based on the number of tasks they have successfully completed (typically, in excess of 5,000) and their according acceptance rate (typically, 95% or better). With these types of projects, looping is not required as the level of trust in a master worker is justifiably high enough to accept their judgment-based answers as correct, especially when their accuracy is measured against known results. These types of project are not a social phenomenon, but rather more like the traditional model of hiring skilled consultants to complete a job. The use of crowdsourcing makes this traditional model work far more efficiently than in the past.
Does “gamification” make crowdsourcing social?
Taking mundane tasks and turning them into “games” in order to increase output has been going on in some form or fashion for decades. Both social sites and crowdsourced projects have embraced gamification as a means of increasing quality and production volume. Gamification does not, however, make the crowd any more “social” as a result, mainly because the “games” that accomplish a specific task are simple and repetitive, a far cry from the elaborate, skill-testing activities over which gamers interact with each other. Gamification works both ways, and the crowd can quickly turn mischievous or even deliberately devious by exploiting vulnerabilities in a process for fun or for personal financial gain.
So…is the crowd social?
Yeah, but no. Crowd workers do communicate with each using social media, much in the same way that fans of a particular movie franchise or people in any particular niche would communicate directly with each other. The use of “fans” in social media is truly a social phenomenon, but can only loosely be defined as crowdsourcing. The employment of master workers is akin to hiring highly-qualified consultants. But the looping phenomenon is absolutely a social-based method of data collection. The use of multiple workers to find data that the machine—an uncaring, unbiased third party—agrees upon as correct is absolutely social.