Facebook is in the news more than ever at the moment, as its CEO Mark Zuckerberg fights allegations of voter manipulation following a New York Times investigation into the 2016 US election.
Behind the media portrayal of Facebook as a social network aimed only at connecting people, are its many tech patents. These show how the company is looking for ever more sophisticated ways to utilise user data for advertising purposes.
Targeting extended family networks
On 15 November, its latest patent application was published, showcasing a system that would make it easier for Facebook to target families and extended networks of people, whether they are on the platform or not.
Filed at the US Patent and Trademark Office on 10 May, the application includes an algorithm that could be used to identify different clues in photos. These could be faces, shapes, brands or other specifics. Facebook aims to use this data to build a profile of the entire household, based on the user’s activity. This would allow them to target a wider range of advertising.
If the company launches the system, it would be an addition to last year’s programme aimed at targeting families. This allows the platform to analyse user data to work out the household members by analysing surnames, location check-ins, event check-ins, relationship listings and shared life events.
Sophisticated algorithms
The recent patent application would take the premise of this system much further. Sophisticated levels of data mining would include cross-referencing photos that tag the user, how many people are positing on Facebook from specific IP addresses, descriptions typed by the user and more. They will use this to accurately guess the number of household members, and their demographic. This would be based on their age, socio economic status and other linked clues.
All of this means that Facebook will know who lives with a user, whether they are on the platform or not. An example given by the social network describes a male Facebook user who posts many pictures that include two women. The women appear regularly either in his own posted photos or pictures that friends tag him in. On one of the pictures of a young girl, the man has typed a description of ‘my angel’. Facebook’s algorithms would use this information to deduce that the household is made up of three people: the man and two women. It would further deduce that one of the women is most likely his wife, and the other his daughter.
Whether this deduction is 100% accurate doesn’t matter hugely for advertising purposes. It would still allow a wider range of advertisers to target this user with goods and services that are aimed not only the male user, but also his probably family members. Even if his wife and daughter do not have an active Facebook profile, this technology allows advertisers to target them through the male user.
Broad claims for new system
As the patent’s claims are broad, it’s tricky to know exactly how specific the system plans to be. It could involve cross-referencing specific facial features or encompass a wider level of data by analysing photographs for subtle clues.
A separate patent filed earlier this year by Facebook covered using phone microphones to listen in on users’ TV choices. Published in June 2018, this patent shows just how many angles Facebook is hoping to cover in collecting user data.
The latest patent talks about ‘demographic composition’ but doesn’t go into any details regarding what this would mean for the user. It covers targeting users based on the projected size of their household, the ages and status of the members and any shared interests.
Marketed at families
Facebook’s advertising firmly advertises how it wants to be seen as a platform that brings families together. The Portal home video chat service launched earlier this year, says that it’s aimed at connecting generations. It includes a microphone that listens 24 hours a day, along with a ‘smart camera’ that uses advanced AI tracking to identify users. There is no evidence that this uses facial recognition tools, however, which are covered in the newest patent.
For those worried about data intrusion, the patent application makes it clear that while it uses technology to analyse photographs posted publicly on Facebook, it won’t delve into any content in videos or private messages. It’s also worth noting that Facebook often files patents that are yet to turn into reality. So, while Facebook most likely could use your birthday photos shared by friends to target you with Christmas adverts, it won’t necessarily do so.
About Dawn Ellmore Employment
Dawn Ellmore Employment was incorporated in 1995 and is a market leader in intellectual property and legal recruitment.
