LinkedIn now ranks dwell time and saves over reactions

A post that racks up 400 likes in an hour and nothing else now performs worse, on average, than one with 40 likes and eleven considered comments. That is the practical consequence of what Sendible Insights calls the Depth Score: LinkedIn's 2026 ranking model, which weighs dwell time, comment quality, and saves above the reaction count that has driven feed strategy since 2018. Sendible describes it as the platform's biggest distribution shift since it abandoned the chronological feed. That claim is not hyperbole. It inverts eight years of received wisdom about what a good LinkedIn post looks like.

The reaction economy is over

The old game rewarded velocity. Post at 8am, harvest likes in the golden hour, let the algorithm mistake enthusiasm for value. Agencies built entire posting calendars around it: short hooks, single-line paragraphs, engagement-bait questions ("Agree?") designed to trigger a thumb movement that cost the reader nothing. It worked because LinkedIn's ranking signals could not easily tell the difference between a reaction and a read.

Depth Score closes that gap. Dwell time means LinkedIn now has a reasonable proxy for whether someone actually read the post or scrolled past it while tapping "like" out of professional courtesy. Comment quality means a ranking system that can apparently distinguish "Great insight!" from a genuine rebuttal or a question that adds information. Saves measure something reactions never could: whether the reader thought the post was worth finding again. None of these signals can be gamed by a pod of colleagues liking each other's posts within ninety seconds of publication, which has been the default engagement tactic for half of B2B LinkedIn since roughly 2019.

Why the format war is really an authority war

This is not a technical footnote. It resets who gets to look credible on the platform. A reaction-optimised feed favoured confident generalists: the executive who could produce a punchy one-liner about "leadership" every Tuesday and watch the likes accumulate regardless of whether the sentence contained an idea. A dwell-time-and-saves feed favours the opposite temperament: people who write something specific enough that a reader stops scrolling, and useful enough that they bookmark it for a client call next week.

That reordering matters most for the institutions with the least tolerance for looking glib. A central bank economist or a multilateral's policy lead cannot compete on wit or hot takes, and should not try; they compete on the fact that a three-paragraph explanation of a capital adequacy rule change is exactly the kind of content someone in compliance saves for later reference. Industrial groups explaining a supply-chain shift, or a philanthropic foundation unpacking a funding methodology, have always had material with genuine save-worthy value. They were simply penalised for years by a system that valued brevity over substance. Depth Score is the first LinkedIn ranking change in years that plays to the natural strengths of subject-matter authority rather than against them.

The corollary is uncomfortable for brand teams that have spent budget optimising for reach metrics that Depth Score now treats as noise. Follower counts and like totals were always vanity proxies; they are about to become actively misleading ones, because a highly liked post with low dwell time will now be suppressed relative to a quieter post that holds attention. Reporting up the chain on "engagement" without distinguishing a save from a reflex-like is going to produce badly wrong conclusions about which content, and which executives, are actually building reach.

What changes on the page, not just in the algorithm

The format implications are concrete. Posts need enough substance to survive past the three-second scroll test: a specific data point, a named source, an argument with a conclusion, not an observation with a hashtag. Comment strategy shifts from "reply to everything fast" to "ask a question in the post that only an informed reader can answer well," because a thread of substantive replies now carries more ranking weight than fifty low-effort ones. And the save button, long the most ignored piece of UI on the platform, becomes the metric worth designing for: content structured as a reference (a framework, a checklist, a explained regulation) earns saves in a way that opinion does not.

For financial-services, multilateral, and industrial communications teams, the practical instruction is blunt: stop measuring campaign success by reactions, and stop briefing executives to write for the like button. Depth Score rewards the institution willing to publish something a reader needs twice.