Reach is down. Engagement is shifting. And we’re probably measuring the wrong thing.
Every few months, the same topic reappears. Reach on social media is down. Algorithms are suppressing organic content. Nobody's seeing anything anymore. The feeds are broken.
And sometimes, the data agrees. Sometimes, it doesn't. What it almost never does is tell a simple story - because marketing measurement has never been simple!
Two new studies put some useful numbers behind the social media measurement challenge. Buffer analysed over 52 million posts across 10 platforms. Metricool looked into 673,658 LinkedIn posts from more than 63,000 accounts. Together, they complicate the narrative in the most useful way possible.
What ‘success looks like’ has changed
Yes, some platforms are showing declining engagement rates. Instagram's median engagement rate fell from around 7.4% in 2024 to around 5.5% in 2025 - a 26% decline.
But here's what that number doesn't tell you: Instagram has increasingly steered users toward views as its primary success metric, which means the traditional engagement rate formula may simply be measuring less of what Instagram is actually optimising for – it was designed for a feed of static images. It doesn't capture saves ("I want to come back to this") or sends ("I want someone else to see this"), neither of which shows up in a public count. The metric didn't break. The platform moved on, and the formula didn't follow.
Meanwhile, Facebook's median engagement rate rose to around 5.6% in 2025 (up from around 5.0% in 2024), Pinterest climbed 23%, and X, despite remaining at the bottom of the engagement-rate rankings, saw a 44% relative gain.
So, is social media performance dying? Not across the board. What's really happening is that each platform is evolving its own definition of what success looks like - and the core metrics we've been using for years aren't keeping pace.
"Engagement" isn't one thing
This is the part most reporting gets wrong. When we talk about engagement rates across platforms as if they're comparable, we're mixing up fundamentally different measurements.
LinkedIn, for example, includes clicks in its engagement rate, while most other platforms don't. The Metricool LinkedIn data highlights this issue; look at the year-over-year numbers for LinkedIn Company Pages: impressions down 10%, likes down 13%, comments down 17%, shares down 11%. On the surface, it looks like a platform losing steam. And if those were your only metrics, you'd probably be worried.
But clicks, the interactions that don't show up publicly on your posts, rose by 5%. Every time someone swipes through a carousel, watches a video, or follows a link, LinkedIn tracks it. When you include those invisible interactions, overall engagement rate went up, from 12.21% to 13.90%.
LinkedIn now surfaces impression metrics for comments, meaning the conversation itself is being measured for reach. And posts don't just generate engagement; they drive profile views and follower conversions that don't register as "engagement" in most reporting dashboards, but are often the actual business outcome the post was trying to achieve.
A post that gets 200 likes and converts 40 profile visitors into followers is doing more useful work than a post that gets 500 likes and sends nobody anywhere. Most analytics tools report the first number. Virtually none report the second.
People are interacting with LinkedIn content more than ever. Just not in ways anyone else can see. And that's not just a LinkedIn story.
Instagram’s most valuable engagement signals — saves and sends — are invisible to everyone except the creator and the algorithm. A save tells Instagram that this content has enough value that someone wants to return to it. A send tells Instagram that this content is worth someone's social capital to share privately.
These are higher-intent signals than a like, and they're exactly the ones the traditional engagement rate formula was never built to capture.
And then there’s the role of comments; not just as engagement, but as content in their own right. The original post is often only the starting point; the real narrative, social proof, and persuasion increasingly unfold in the comment thread.
Across nearly two million posts from 220,000+ accounts on Threads, LinkedIn, Instagram, Facebook, X, and Bluesky, posts where creators reply to comments consistently outperformed those where they don't — on every platform studied.
The estimated engagement lift: Threads +42%, LinkedIn +30%, Instagram +21%, Facebook +9%, X +8%, Bluesky +5%.
The Like button: a great 2009 answer. Just not a 2026 one.
The Like button turned 17 this year. When it launched in 2009, there were no carousels, no Reels, no Stories. The feed was a chronological stream of status updates and photos. Engagement meant just one of two things: you pressed a button, or you left a comment.
Now, swiping through a carousel is engagement. Saving a post to return to later is engagement. Sending something to a colleague with a note is engagement. Clicking through to someone's profile after reading their post is engagement. None of these behaviours fit into the original model, and yet they're now among the highest-intent signals a platform receives about whether content is actually working.
The platforms have been quietly reweighting their algorithms around these user behaviours for years. What's changed is that the gap between what the algorithms are measuring and what most social media managers are reporting. The conversation has not moved on from asking about followers and likes.
The format gap
Both studies independently arrived at the same finding: the formats that dominate in posting volume are almost never the formats that perform best.
On LinkedIn, images account for around 49% of all Company Page posts, and video another 25% — so nearly 75% of all content. Yet carousels, which make up just 7.6% of posts, earn a 49.52% engagement rate. Images earn 5.77%. Video earns 6.91%. Metricool's headline finding: carousels get 11x more interactions than images, yet images are posted 6x more than carousels.
Metricool also found that LinkedIn posts including a direct question see 77% more comments than average, and posts with a specific call to action to comment see an 80% increase. That's not a hack, it's simply designing content to invite a response rather than passive consumption, and measuring whether the conversation happened, not just whether the post was seen.
There’s a similar story on Instagram — and it’s really a reach story. Reels get 36% more reach than carousels, putting your content in front of people who’ve never heard of you. But carousels earn 109% more engagement per person reached, meaning the people who do see them are going much further in. These aren’t competing findings; they’re pointing to two different strategies for two different goals. Reels are for reach — discovery, new audiences, the top of the funnel. Carousels are for depth — getting the people who already follow you to go further in. The mistake is treating “best format” as a single answer when the platform is really asking: best for what?
And top tip, add music to your carousel and Instagram will treat it as a Reel.
And it's worth mentioning polls on LinkedIn, which reach nearly three times more people per post than any other format and are simultaneously the least-used format of all.
Part of the tension here is operational: we're often not creating content that aligns with the engagement outcomes we're actually worried about. Instead, we're defaulting to what is easiest, familiar, or historically “standard” within teams.
Reach: the right question
So is reach actually collapsing? The honest answer is: it depends what you’re posting, and why. Organic reach has undeniably become harder on some platforms — LinkedIn impressions for Company Pages fell 10% year-on-year, and algorithm-driven feeds mean fewer posts surface to non-followers by default. That’s real.
But reach is not evenly distributed — it’s format-dependent. LinkedIn polls reach nearly three times more people per post than any other format on the platform, yet they’re the least-used format of all. On Instagram, Reels consistently outperform carousels on raw reach. The platforms aren’t suppressing content uniformly; they’re rewarding formats they want to promote. Knowing which formats those are is the more useful question than asking whether reach is “up or down” overall.
And here’s the more fundamental point: reach was never the destination. It was always a proxy — a way of estimating whether the right people might have seen something. If your content is generating the outcomes you actually care about (pipeline, profile growth, inbound enquiries, qualified conversations), then lower headline reach numbers are almost irrelevant. A post seen by 500 of exactly the right people that drives three meaningful conversations is doing more useful work than a post seen by 50,000 people who scroll straight past. The question worth asking isn’t “how do I get more reach?” — it’s “is my reach reaching the right people, and are they doing anything as a result?”
The boring truth about consistency
In the analysis of 4.8 million observations across approximately 161,000 profiles on Facebook, Instagram, and X, Buffer found that accounts that went quiet for a week consistently underperformed their own baseline growth rates. They called it the "no-post penalty."
The headline isn't that you need to post more. It's that going quiet has a penalty.
What engagement is really becoming
The shift from public to behind-the-scenes engagement isn't a temporary algorithm quirk. It reflects something real about how people use these platforms now — more intentional, more private, more likely to save something for later or send it to one specific person than to perform a public reaction to it.
For individuals and organisations, the metrics worth building dashboards around are increasingly the ones that don't show up in the public domain.
And underneath all of it, the findings from both studies point to the same foundational behaviours: show up consistently, use the formats that generate intentional engagement, and be social on social.
And when someone asks whether your reach or engagement is down — ask them what happened next. Because if the right people saw it, engaged with it in ways that mattered, and something moved as a result, those are not metrics problems. That's the whole point.
Sources: Buffer, State of Social Media Engagement 2026; Metricool, LinkedIn Study 2026
