A Facebook engineer has created technology that overwrites set policies in the company’s platform to increase the likelihood of the ad being included in the news feed.

Based on
a likelihood of the user interacting with an advertisement, the online system may modify one or more of the advertising policies to increase a likelihood of the advertisement
being included in the feed.  

The United States Patent and Trademark Office granted Yi Tang, engineering director at Facebook, a patent last week. The patent,
which related generally to the presentation of content by an online system, was originally filed in December 2014.

The abstract explains that an “advertising policy specifies a
minimum distance between advertisements in the feed. Based on a likelihood of the user interacting with an advertisement, the online system may modify one or more of the advertising policies
to increase a likelihood of the advertisement being included in the feed.”

The patent enables the advertising system to ignore the policy when selecting content for the feed or may
modify conditions limiting inclusion of advertisements in the feed specified by the advertising policy.

The patent explains that when Facebook receives an interaction with content
from a user, the online system stores information describing the interaction and may generate content describing the interaction.

Presenting users with content related to the interaction may
increase user interaction with the online system. This situation aims to “persuade” online users to take an action such as a clicking on a product ad or video. Many online technology
platforms may receive compensation for that user taking action.

As users likely interact with the advertisements, including organic content, enforcing advertising policies may prevent the user
from being presented with content that the user is likely to interact with.

To increase the likelihood of a user interacting with a feed of content, the patent describes the online system as having the ability to modify one or more advertising
policies based on the likelihood of the user interacting with advertisements from various ad requests.