An analysis of the 200 most-recent tweets on X by 18 big publishers indicates links seem to hurt engagement, after debate arose over X's value for sharing news
Engagement for tweets from @nytimes (53 million followers) is dwarfed by engagement for tweets from @GlobeEyeNews (866,000 followers).
Nieman LabLaura Hazard Owen
Context & Ripple Effects
X’s usefulness as a distribution channel for publishers was already weakening before this engagement analysis: Chartbeat’s publisher data showed Twitter’s referral share shrinking over several years, while Similarweb also recorded a subsequent monthly decline across major outlets. The new finding adds a format-level explanation for why publishers may struggle to convert social reach into visits.
The 18 publishers have a direct incentive to test fewer link-led posts and more native X formats if their immediate goal is on-platform engagement; that tradeoff can reduce the traffic value of each social post.
Large publisher accounts, including @nytimes, cannot assume scale translates into engagement when smaller accounts can outperform them on interactions despite far smaller audiences.
Second-order effects
Social teams will have to distinguish engagement optimization from audience acquisition: declining referrals and lower-performing link posts make X harder to justify as a dependable route to owned publisher sites.
If link suppression is systemic rather than publisher-specific, competing outlets face the same pressure to package news for the platform, potentially rewarding posts designed for reactions over click-through context.
Third-order effects
The pattern would further shift publisher-platform relations from referral partnerships toward rented-audience distribution: publishers retain the production cost while the platform captures more consumption and engagement.
It may also deepen the industry’s measurement problem. Where engagement is easier to earn than qualified traffic—and has previously favored unreliable sources—news organizations may need to weight conversion and trust outcomes more heavily than interaction totals.
The trend: This is one data point in the broader shift from social platforms as meaningful publisher-referral channels to platforms that prioritize native attention over outbound distribution.
@TheStalwart I mean, there's really 4-5 things: 1) Per se algorithmic penalty for links (disputed) 2) UX “enhancements” that discourage clicking links 3) Deprioritization of “following"/chronological feed in favor of “for you” 4) Learned behavior from users #3 probably the most i…
The NYT published a link to critical original reporting on Iran 45 minutes ago. A good, fair story. They have 53m followers. The engagement metrics you display say they got 94 likes and 33 retweets out of that. Is that accurate? And if so, shouldn't you work on a better algo? [im…
In what sense is it intellectually dishonest, Nikita? The context is clear. My thesis is: I think your algo surfaces too little quality. And I think your excuses in these back-and-forth exchanges have been validating of that. To your credit, it's clear enough you agree with me.
@nikitabier ... Here's a crazy thought: What if you just showed tweets to the accounts that follow them? The whole point of Twitter is that I control my feed, not you. Wild that you don't understand that.
After the exchange between @NateSilver538 and @nikitabier, I did a little test to check whether that was true and, to my surprise, what I found suggests that link deboosting was indeed reversed. What I did is randomly sample 15 tweets by @nytimes between 2019 and today, compute […
@nikitabier ... Obviously something in the algorithm changed in 2023, because clicking on a link was also covering the engagement buttons before that, yet the reach of posts containing a link suddenly collapsed around that time. This may not have resulted from an explicit penalty…
@phl43 ... Just to be absolutely crystal clear: 1. Links were never “deboosted” 2. They received lower engagement because websites cover the very buttons that cause something to rank better in the algorithm. 3. Today, the buttons are no longer covered, so links now have the same …
People like me who have been using Twitter through every iteration have a lot of muscle memory about what drives traffic and ultimately revenue. It used to be much easier to get people off site and that created *more* incentive to build your following by engaging on-site too.
@TheStalwart If you follow an account like @nytimes that doesn't “engage"/interact much b/c you're using Twitter as a news feed to “monitor the situation”, and there's little algorithmic reward for following an account and none for clicking on external links, it will get depriori…
It's almost certainly true that part of the reason the NY Times twitter account doesn't get much engagement is because they're not really even trying. But also: [image]
FWIW the NYT doesn't get much engagement because we changed the rules for how links worked and they didn't play ball because they're a news organization that publishes reporting on their own website and not ours. — (via www.niemanlab.org/2026/04/do- l...) [image]