What actually happened
LinkedIn rolled out the Depth Score ranking signal across feed distribution this spring. The platform now weighs how long a viewer actually engages with content, measured in seconds of dwell, more heavily than likes, comments, or reposts. A user who watches your video for 45 seconds and never reacts now signals more value to the algorithm than a user who taps "like" in 1.5 seconds and scrolls. The hierarchy of signals changed.
The same update knocked company-page organic reach down roughly 60 to 66 percent between 2024 and 2026, and pushed company-page allocation to about 5 percent of the average user's feed while personal profiles take roughly 65 percent. Native video continues to outperform static and text posts, but the gap that matters is no longer engagement rate. It is finish rate. Reach lift is now tied to viewers who stay, not viewers who tap.
And the timing matters. LinkedIn passed YouTube this spring as the top B2B video destination by marketing-buyer survey share. The biggest channel for B2B video just changed how it ranks B2B video.
Why this rewires the playbook
For two years, the dominant advice for B2B video on LinkedIn was: copy TikTok. Lead with a 3-second pattern-interrupt. Stack curiosity gaps. End on an open loop. The whole craft was tuned to a thumb-stop.
That playbook now optimizes for the wrong metric. A scroll-stop gives you a fraction of a second of dwell, and if the next 10 seconds don't hold, the algorithm logs it as a low-Depth-Score finish and throttles future distribution. The pattern-interrupt opener you have been training your founder on actively trains the algorithm to deprioritize your channel.
What wins now is structure. A defined teach, an obvious arc, a viewer who understands within five seconds why staying another 60 is worth it. Calm tone. One idea. Slow enough to follow on 1.0x speed.
That happens to be what a content system already produces. It does not happen to be what a $400 freelance edit with a "punchy hook" produces. The market has been quietly rewarding the wrong craft for two years. The reward is over.
The data
From our retainer book. Across our 2026 client engagements we have been splitting LinkedIn-native cuts into two buckets and tracking finish rate against impression lift week-over-week.
The hook-stack edits — heavy pattern-interrupt opener, fast cuts, captioned curiosity gap — show a high one-second view rate. Average finish rate sits in the high 20-percent range. Impression growth flatlines or decays after week three. These are the videos that look like they are winning on a Tuesday-morning report and have stopped getting reach by the following Tuesday.
The structured teach edits — calm establishing shot, named promise within five seconds, single thesis, 60-to-90 seconds — show a lower one-second view rate, but finish rate runs in the 50-to-60 percent band on most clients. Impression growth compounds. The second post in a series typically outperforms the first by 1.5x to 2.5x. We have stopped quoting first-second view rate to clients. Finish rate at 45 seconds is the only number that matters now.
From third parties. LinkBoost's late-2025 algorithm teardown puts the company-page reach drop at 60 to 66 percent between 2024 and 2026 and confirms LinkedIn's internal weighting now penalizes posts with external links by roughly 60 percent (linkboost.co). Sprout Social's 2026 piece on the broader social-video shift shows the 60-to-90-second range outperforming the 7-to-15-second range for B2B-coded creators across platforms — the format shift is not LinkedIn-only (sproutsocial.com). And b2the7's May 11 trend report frames the shift bluntly: LinkedIn is now the top B2B video platform and depth, not flash, is its new floor (b2the7.com).
The counter-argument, steelmanned
The strongest case against this thesis goes like this. Dwell time has always mattered. Hooks are not the enemy. A good hook is exactly what gets people to stay. You are describing bad hooks, not the craft of opening a video.
That is fair. A 3-second opener that promises something specific and then delivers it inside the next 87 seconds is not the problem. Pattern-interrupts that earn the next 60 seconds of attention still work.
The argument breaks where the real market lives. In practice, the hook-bait playbook produced two things: openers that promised more than the video could pay off, and openers built to win the thumb-stop rather than the finish. Both pull average dwell down. We have audited enough B2B feeds at this point to be confident the average marketing team is not producing the rare "high-tension hook plus rigorous payoff" video. They are producing the common "good opener, weak middle" video. The Depth Score signal punishes that modal case harshly.
The other counter is timing. LinkedIn changes are reversible. Maybe. But the underlying force — platforms shifting their economic model from impressions to attention — is not. Spotify is paying podcasts on listen completion. YouTube has been weighting watch time for a decade. TikTok now boosts the 60-to-90-second range for creators who hold attention. The Depth Score is not a fad. It is the format catching up to where the rest of the attention economy already was.
What to do Monday
First, change the metric in your weekly marketing report. Replace "average reach per post" with median dwell at 45 seconds. If your team cannot pull that out of LinkedIn Analytics directly, calculate it from finish rate multiplied by duration. The number is the number, and your retainer or in-house team should be reporting on it by next Friday.
Second, audit the last 10 LinkedIn videos your team posted. For each one, write down the actual promise the first five seconds makes. If the promise is "this will be surprising" rather than "this will teach you X," it is hook-bait. Cut it.
Third, run one experiment this week. Take a piece of content you would normally edit as a 30-second cut with a hook, and post it instead as a 75-second structured teach with the thesis named at the five-second mark. Track finish rate at 45 seconds. Repeat on the next two posts and compare against the hook-stack baseline.
Fourth, shift posting weight from the company page to personal profiles. The 60-to-66-percent company-page reach drop is structural, not seasonal. Your CEO's profile is now your distribution channel. Treat it like one — production schedule, talking-points pipeline, captioning standard, the works.
Fifth, stop producing 30-second cuts as the primary deliverable for LinkedIn. Make them downstream artifacts. The 60-to-90-second structured teach is the new primary asset. The 30-second cut is the cutdown. Reordering your asset hierarchy is the cheapest, fastest change you will make this quarter, and it has the largest knock-on effect on what your production partner is actually building for you.
The operator POV
If you are buying B2B video on a retainer, the Depth Score change is good news. The format the feed now rewards is structurally easier for a content system to produce than for a freelance editor armed with TikTok tropes. The discipline that wins is structure, not stunts.
If you are buying B2B video as one-off projects, this is the second algorithm shift in 18 months that has made one-offs a bad asset class. Each new shoot has to relearn what dwell looks like for your audience. A system that has watched 12 weeks of finish-rate data will out-iterate the freelance edit every time.
Either way, the next 90 days of LinkedIn distribution will be decided by which marketing teams retrain their craft fastest. The teams still chasing one-second view rate are going to find out the hard way that they have been training the algorithm to bury them.