How we infer demographics
We estimate or infer demographic characteristics of local consumers based upon their social media presence/usage, including their public posts, the accounts/pages/posts they follow or like, and the people who follow/like their own posts. Our methodology is data-centric, relying on multiple data signals from three primary areas: networks, consumption and language. Data signals are filtered and amplified using large, proprietary knowledge bases of established correlations between data points and demographic characteristics. Finally, we combine multiple amplified signals using a series of algorithms to infer likely demographic characteristics. The specific demographic characteristics that we infer include gender, age, marital status, parental status, location (never more granular than country/state/city), ethnicity, religion, political affiliation, occupation and likes/interests.
Our inferred demographics are estimates based purely on social media presence/usage. We never cross-match or combine in any way with actual demographics from third party databases or any other sources. Information we collect in relation to usage of our own products, services and websites is likewise never cross-matched or combined in any way with our inferred demographics.
How we use inferred demographics
We use our inferred demographics for products and services that give our customers and partners insight into the consumers who follow their social media accounts, post about their brands, and respond to their social media marketing campaigns. We also use inferred demographics to help customers and partners target their messages and advertisements more effectively: in this case, we use systems that facilitate targeting without revealing the inferred demographics of individual identified consumers.
Delivery via our demographics API
When our inferred demographics are supplied to customers or partners via our demographics API, they are either aggregated or anonymized. Aggregation means that customers or partners receive aggregate reports for groups of individuals, based on our inferred demographics (for example, we might report that 57% of the followers of a certain social media account have been inferred as female). Anonymization means that customers or partners receive our inferred demographics for individual consumers, but these exclude individual identifiers, such as names or usernames.
Requests to delete inferred demographics
Any individual may request that demographics we’ve inferred based on their social media behavior/usage be permanently deleted. Requests via email to firstname.lastname@example.org, referencing the social media username of the individual, will typically be processed within one working day.
Effective date: 14th September, 2014