Instagram's head measures the cost of its own off switch at half of engagement

The person who actually finds the setting has heard the feed is deciding things for her, so she digs into the menu and turns ranking off. A week later she is looking at accounts that post several times a day, and the friend who posts once a month is nowhere. Mosseri described that mechanism accurately when he said a chronological feed ends up overwhelmed with content from companies, brands and publishers ^[3].

It is also a problem Instagram built. Professional accounts post at volume because the ranked feed paid them to, and switching the ranking off does not untrain the behaviour, as thenextweb argues in its account of the tradeoff ^[13].

Take the 50% at face value and it is a statement about the product's economics: it takes two opted-out users to produce the engagement of one ranked user ^[14]. Mosseri offered engagement, time spent and satisfaction. He did not offer any measure of wellbeing or of the regret people report afterwards ^[10]. Satisfaction is the closest thing on his list to a user-side number, and it is satisfaction with the feed.

Meta has been telling a different story with the same variable. The company has presented its AI increasing time spent on Instagram as a success ^[6], and Mosseri now presents less time spent as something users lose ^[4]. Both readings hold together only if time on Instagram is good for the person spending it, which is the proposition a jury in the Meta social media trial is being asked to weigh ^[7]. On whether the app is addictive, Mosseri said the claim was controversial and separated colloquial from clinical use of the word ^[12].

The prompt in the Australian draft targets the much larger group who never open settings, not the woman in the first paragraph, who already found it, which is why the draft requires prompting rather than availability ^[16]. Instagram has already run the availability version: the chronological option it restored in 2022 had to be selected each time rather than staying on, so the default returned whenever the app did ^[9].

What an off switch is worth comes down to two things: who opens the conversation, the user or the product, and whether the choice survives the next session. Instagram's 2022 build sat in the corner where the user hunts and nothing persists, which reliably produces a small adoption number and the argument that nobody wanted it. Article 38 of the Digital Services Act has required very large platforms to offer a non-profiling feed since 2023, and the modest uptake has been read by the industry as proof of disinterest ^[11]. Mosseri's figure gives that reading a competitor: if opting out halves engagement, a company has an arithmetic reason to leave the switch somewhere few people trip over it ^[17].

For anyone shipping a control that competes with their own default, the number worth having is the gap between how many users change it when asked once and how many find it unasked. Where those two figures are far apart, the adoption rate was measuring placement.