The Facebook Algorithm Identifies Family Members To Provide Ads

A Facebook algorithm has been recently introduced to analyze family photos to target ads. Its new patent identifies the relationship among members of a family, even when not all of those have a Facebook account.

Through this patent, Facebook won’t have many difficulties in analyzing family photos and targeting ads. This Facebook algorithm was submitted on May 10^th and officially announced on November 15^th.

What makes the patent special is that it can analyze any details in a photo like faces or pets, and compare them with other data available in Facebook’s data store to classify and identify members in a family.

We can easily see that Facebook has taken a big step in supporting advertisers to recognize their potential customers. The company has expanded to targeting entire households instead of targeting individuals as before.

The first time Facebook revealed about a programme to target family members in photos was one year ago, 2017.

The Facebook algorithm will exploit abundant data sources available including users’ profiles, life events, similar last names or check-ins and so on to build a corresponding family profile. That means Facebook won’t only see photos, but also collect all of the possible information relating to users’ profiles.

The algorithm helps Facebook to determine members in a family by searching for IP addresses. Those who are sharing the same IP addresses may be a family. Besides that, it can look for descriptions and tags.

Users’ private life will be no longer a secret to Facebook when the social media giant can know everything from the number of members in their family, their lifestyle and much other information.

The accuracy of the Facebook algorithm may not reach 100%; however, it is enough for advertisers to look for its potential customers as Facebook users.

According to Facebook, the patent mines only public photos, not contents or videos on private messages. That means users can still keep secret for their private data. Additionally, it may be an unused feature and a patent is just a patent, no more.