May 6, 2019
Companies’ main reason for collecting a lot of data about users is to improve targeting of ads, which allows them to make more profit from advertising. But big data need to be analysed correctly. The Pew Research Center polled near 1000 US Facebook users aged 18+ about understanding of the site’s categorisation algorithm () and found that 74% of respondents did not know that Facebook maintains a list of users’ int...erests, hobbies and preferences on their ‘Your ad preferences’ page. Although 59% said that categories listed reflected their real-life interests, 27% of users said the categories were not very or not at all accurate in describing them. A second Pew survey of US social media users (Facebook, Twitter, Instagram, etc) found that most people believed that it would be relatively easy to determine their race or ethnicity (84%), hobbies and interests (79%), political affiliation (71%) and religious beliefs (65%) from But for 11% of respondents conclusion was “You have no behaviours” (not assigned to a category). Other participants were rather uniformly distributed by quantity of categories fewer then 10 categories (27%), 10-20 (27%) and +21 (33%). Facebook was a bit better at categorising users’ political views: 51% of users were given a political label and 73% of them said that this described their views very accurately (31%) or somewhat accurately (42%). Half the users (51%) were not comfortable that the company had created such list.