Many applications and companies are currently using the potential of Big Data Analytics in different ways. Smart cities, customer segmentation, or personal and public health are just some of the use cases. Data are out there, online, stored in databases, accessible through APIs, social networks,… The race of applications using complex algorithms to analyze the data has only just got off the ground. But, what happen if data we are using are not the adequate for the expected result? We would be building on quicksand.
For instance, brands and companies want to be customer-centric, and therefore there are applications for customer analysis, prospects detection, marketing campaigns design, etc. The vast majority of these applications uses information available on social networks such as LinkedIn, Facebook or Twitter. They use this information to analyze and target potential customers. But there is a problem: where are they extracting the information from? Do all social network users publish two posts a day (because of their sharing policy)?, Are their posts showing their interests / needs or just their own personal digital marketing?
The fact is that users take different roles in social networks. If our algorithm draws relevant conclusions from the comments made on social networks, we are leaving out the vast majority of the contacts ―that do not use to post on social media but participate in another way. 29% of social media users publishes 85% of the content of the social networks. Retrieved information, processed in this manner, is not enough. Social network users do not completely represent potential customers, in the same way that the people who overshare do not represent the total of your contacts. For those reasons, new mechanisms for analyzing the Web are necessary, so we can draw conclusions including the participation of other users.
In Relevante.me we are aware of this challenge and we are working on new methods to analyze social information.