Using a mixture of gamification and crowdsourcing, local online reputation management (ORM) agency BrandsEye is looking to significantly improve the accuracy of gauging online sentiment.
BrandsEye Crowd is a newly-developed tool that draws on a large community of people from around the world to bring local context and judgement when it comes to analysing online mentions.
According to BrandsEye MD Tim Shier, while crowd sourcing is not new, BrandsEye is the first company globally to adopt such principles to tackle accuracy in the online reputation monitoring industry.
“[BrandsEye Crowd] is a radical innovation in the online reputation management space. We know of crowd-sourced rating in academia, especially economics, but not for commercial brands. We're very proud of the system, which is already showing incredible accuracy, and we believe can get to between 95-100% perfect with a bit more refinement. In this space, accuracy is all-important,” says Shier, adding that the typical margin of error in ORM data is as high as 20%.
“Computers are not great at picking up irony, sarcasm, humour or rage. People are,” says Shier, explaining that people looking to earn some extra money can sign up to become part of the BrandsEye Crowd as “raters”.
Peer review
“The rater gauges if each mention is relevant and judges the sentiment, the location of the person that made the mention and the media type. Local context and local language subtleties are critical - along with the intuitive sense humans have for what other humans are trying to say.”
BrandsEye has clients in 113 countries, and will be implementing the system globally, using local people who understand the local context, language and nuances. “For example, in SA some terms have different meanings to what they may have elsewhere in the world, and this is key when brands need to do strategic planning.”
Shier explains that as the data comes through, two people are fed the information and rate the data according to the system's criteria. “If those two people agree, then it is deemed correct, if they disagree, it is then sent to a third person for rating. If at that point they don't agree then our other mechanisms kick in.”
The gamification element comes into play in the way in which raters can build up their accuracy scores and earn more compensation as a result. Shier says the Crowd is backed by a sophisticated computer platform that uses the same technique as that used for ranking chess grandmasters, in order to effectively rate the members of the crowd and to ensure accuracy.
Top-class tech
Shier says BrandEye is looking to target the underemployed and unemployed (such as stay-at-home mothers and students) to become raters for the service. Raters can earn in the region of R3 000 per month, but this is dependent on their performance. Crowd members can also specify areas of interest and will be assigned data and trained by the system accordingly.
“If someone is performing badly, their accuracy rating will go through the floor and they're not going to earn any money.” In terms of recruitment, Shier says people have already been joining the Crowd through word of mouth only and there is a substantial waiting list. He adds that the size of the Crowd in different regions will be scaled according to demand.
“Our clients can choose how much data they want to see pushed through their account, and we will scale the size of the crowd needed for the analysis accordingly,” says Shier. In the event of a brand crisis or special event, the crowd can also be scaled up almost instantaneously.
BrandEye Crowd has been in development for the past four to five months, and Shier says he is extremely proud of the result.
“I think it's very important that we stop seeing ourselves as the poor African cousins. In SA, we can create technology on par and in many cases better than what the rest of the world has to offer.”
Share