Facebook’s better money-spinner is its News Feed, and the aggregation has appear yet addition key change to the algorithm.

In a fresh announcement, the social arrangement said that it’s leveraging new signals such as user surveys to prioritize posts from close accompany and agreeable you might find worthwhile.

The shift marks Facebook’s connected change from a public belvedere to a more closed-off, claimed space. By absorption on allusive interactions amid people, the social media aggregation seeks to ambit from a string of issues – polarization, fake news, hate speech and agitator agreeable – that have eroded public’s trust in it over the last few years.

Facebook CEO Mark Zuckerberg afresh said the company’s administration would shift from being a social arrangement where people advertisement advice to large groups of people — a town square — to a annual that functions as a digital agnate of the living room, where people acquaint with abate trusted groups.

The News Feed algorithm already factors in indicators like how often you collaborate with a given friend, the number of mutual accompany you share, and whether you mark that person as a close friend.

But by administering surveys to take into annual the accompany and pages whose updates you would absolutely like to see on your News Feed, Facebook appears to be taking a crowdsourcing-like access to rank posts on the feed. This finer means you will see more posts from your abutting buddies as against to abroad acquaintances.

In a agnate vein, Facebook conducted surveys to find out what type of links people collaborate with the most. The social arrangement then accumulated this advice with other capacity about the post – like the type of post (link, video, etc.), the publisher, and the assurance it’s accustomed – to accurately adumbrate if you are likely to find a link valuable.

The change will most likely have an impact on Pages that share clickbait. It follows agnate moves Facebook undertook last month to weed out blah content, misinformation, and other forms of ‘borderline content‘ that brings down “the all-embracing affection of discourse” on its platform.

Of course, employing user surveys sounds a lot like collaborative filtering, which about works by extrapolating similarities amid users’ tastes in order to make new predictions. Recommender systems about rely on such clarification to serve recommendations tailored to your preferences. Netflix and Amazon already use this address so as to advance movie titles and shop for articles (“Customers who bought this item also bought X”).

With over 2 billion active users on its platform, Facebook has become a analytical tool for publishers and users alike. But the tech giant’s profit-at-all-costs agenda has led to a abatement in user assurance at the cost of Instagram (which it owns) and Snapchat.

Read next: Here's how we can change the Earth's orbit to escape the accretion sun