Social Media Analytics: Metrics, Formulas & Benchmarks

LinkedIn's engagement rate is 6.5%. Instagram's is 0.48%. Same creator, same effort, different platform, 13x the difference.

That one comparison tells you more about social media analytics than most 5,000-word guides, because the number isn't a number. It's a decision. A B2B founder spending two hours a day on Instagram carousels when their audience lives on LinkedIn is leaving most of their reach on a platform that won't return it.

I built Sydium after years of bouncing between five dashboards trying to make sense of numbers like these. So here is the frame I wish someone had handed me: every metric you track should sit one link away from a business outcome, and if you can't draw that link, you're tracking decoration. This guide is that frame, plus the formulas and benchmarks to fill it in.

The Only Question That Sorts Good Metrics From Bad

A metric earns its place when a drop in it forces you to change what you'd post tomorrow. That's the whole filter. Engagement rate drops, you fix the hook or the format. Conversion rate drops, you fix the landing page. Follower growth rate drops, you fix positioning. Each points at a lever you can pull. Raw follower count drops and there's nothing to do, because it was never connected to an action. That's the line between a KPI and trivia.

Run every number you track through that question; most won't survive, and that's the point. Access to data was never the bottleneck (59.5% of marketers already use AI for reporting). Knowing which three to act on is. So pick one goal, pick three to five metrics, and ignore the rest:

• Brand awareness: reach, impressions, share of voice.
• Community: engagement rate, reply rate, saves.
• Sales: CTR, conversion rate, referral traffic.
• Growth: follower growth rate, profile visits.

If a metric isn't on the list for your current goal, it doesn't belong in your weekly review.

Metrics Chain To Outcomes In One Direction, And You Should Read Them In That Order

Vanity metrics live at the top of the funnel, outcomes at the bottom, and value rises as you go down. Reach and impressions tell you how many the algorithm showed you to. Engagement tells you how many cared. Clicks tell you how many moved. Conversions tell you how many paid. A number is only as useful as how far down that chain it sits, which is why a post can be a hit and a failure at once: thousands of impressions, zero clicks, the algorithm pleased and the business empty-handed.

Awareness Metrics Measure Who Saw You, Not Who Cared

Reach is unique users who saw your content; one person seeing it three times still counts once. It's smaller than people expect: Instagram organic reach has dropped to 3-4% of followers in 2026, Facebook to 1-2%. At 10,000 followers, 300 to 400 see each post; the rest are theoretical.

Impressions count total views including repeats. Running far above reach, they mean the algorithm is recycling the same audience: sticky content or a small pool.

Views is Instagram's unified 2026 metric, one number across all content types that replaced "Impressions" and "Plays." Any guide still splitting them out is outdated.

Engagement Metrics Measure Who Cared, And They Are Moving Underground

Engagement rate is the most argued-over number in social: at least six formulas, and the one you pick can swing the same data from 0.48% to 6.5%. An engagement rate without a stated formula is unreadable (the six are below).

Saves and sends are the metrics most dashboards still bury. Instagram weights them above likes, so 50 saves can beat 500 likes. A like is acknowledgment; a save is intent, a signal that the content has utility. Optimizing for likes optimizes for 2023.

Shares push your content into someone else's audience. Research shows 44% of people share because they agree, 29% because it's informative, 24% because it's inspirational. Reverse-engineer the reason and you reverse-engineer the post.

Comments take effort, so they signal more than likes, but count lies. Five comments arguing your point beat fifty saying "great post."

Dwell time measures how long someone stops, and LinkedIn weighs it heavily: a 10-second pause pushes a post further than a 2-second one. X runs on quick reactions instead, which is why X-native tactics flop on LinkedIn.

Conversion Metrics Measure Who Moved, And This Is Where Vanity Meets Reality

Click-through rate (CTR) is the first number reach can't fake. Thousands of impressions and almost no clicks isn't engagement, it's decoration:

CTR = (Total Clicks / Total Impressions) x 100

Conversion rate tells you whether the landing page kept the post's promise. People click but don't convert? The problem is usually the page, not the post.

Conversion Rate = (Conversions / Clicks) x 100

Social referral traffic is visitors arriving from your platforms. Tag every link with UTM parameters or Google Analytics dumps social into "direct" and the bottom of your chain goes blind.

Growth Metrics Measure Trajectory, Which Raw Follower Count Hides

Follower growth rate beats follower count at the only job that matters: telling you whether your content is still pulling in new people.

Follower Growth Rate = (Net New Followers / Starting Followers) x 100

A 2% monthly rate at 1,000 followers is healthier than 0.1% at 100,000. The rate reports trajectory; the raw number looks impressive while saying nothing about where you're going.

Six Ways To Calculate Engagement Rate, And The Benchmark Is Useless Without One

Benchmarks disagree across sources because each picks a different formula and rarely says which. Social Insider reports Instagram at 0.48% (by impressions); Buffer reports 4.3% (by reach). Same platform, 9x apart, both correct. Pick the formula first, then read the benchmark.

1. By reach (ERR), most accurate. Purest measure, but needs reach data competitors won't share.

ERR = (Total Engagements / Reach) x 100

2. By followers, best for comparison. Their reach is private; follower count is public.

ER = (Total Engagements / Total Followers) x 100

3. By impressions, best for paid content. Always the lowest, since impressions repeat.

ER = (Total Engagements / Total Impressions) x 100

4. By views, best for short-form video. TikTok and Reels, where the algorithm is view-driven.

ER = (Total Engagements / Total Video Views) x 100

5. Daily engagement rate, best for trends. Long-term patterns and seasonal shifts.

Daily ER = (Total Engagements in a Day / Total Followers) x 100

6. Cost per engagement (CPE), best for ad spend. Paid efficiency or influencer ROI.

CPE = Total Amount Spent / Total Engagements

Each platform defaults to whatever data it exposes: Instagram and Facebook divide by followers, TikTok by views, LinkedIn by impressions. Match the formula to the platform before you compare anything.

The Platform You Pick Decides Your Ceiling Before You Post Anything

Engagement rate varies more between platforms than between your best and worst posts. The 2026 picture, from Buffer's 52M-post study, Social Insider's 70M-post benchmarks, and the platforms' own docs:

Two of those break the rules everyone repeats: Facebook pictures outperform video by 44% even as every network pushes video, and LinkedIn PDF carousels hit 21.77% median engagement, the best of any format anywhere. So repurposing across platforms means transforming the format, not reposting one asset five times.

Your Account Size Bends Every Benchmark, So Stop Borrowing The Wrong One

Smaller accounts beat larger ones on engagement rate, on every platform, every time. It's arithmetic: bigger audiences are less focused and more likely to follow without ever seeing your posts.

Source: Social Insider 2026 benchmarks across 70M posts.

So benchmark against your own tier. A 4% rate at 2,000 followers is 80 people consistently interacting with you, 80 potential customers or collaborators; a 0.3% rate at 500K is more total interactions but a far less invested crowd. The small number is the healthier one, and comparing it to a mega-influencer's is how you quit a strategy that's working.

Replying Is The Highest-Return Metric Move You Are Not Tracking

Buffer's 52-million-post study found that creators who reply to their own comments get an engagement boost on every platform. No new content, no strategy overhaul, just answering.

This is the cheapest lever in the guide, and I'm not theorizing. I grew an audience on X on a reply-first strategy, treating replies as worth far more than likes, and it carried to roughly 332K weekly impressions at the peak. The math is plain: an hour on a new post buys a single-digit engagement rate, an hour replying lifts engagement 21 to 42% and compounds, since more engagement means more reach and more comments to reply to. When you schedule posts in advance, spend the saved time replying, not posting more.

Vanity Metrics Earn A Place Only In Three Situations

A vanity metric looks great in a screenshot and connects to nothing: raw follower count, total likes, page views without context. Run it through the sorting question ("if it dropped, would I act?") and it fails. There are exactly three cases where vanity numbers still earn their keep:

1. Brand awareness campaigns, where reach is the actual goal.
2. Competitor benchmarking, where deeper metrics are private and a public number is all you get.
3. Brand-new accounts, where any signal beats none while you find your footing.

Outside those three, a vanity metric on your dashboard made the cut for emotional reasons, not strategic ones.

Dark Social Is The 50-80% Of Sharing Your Dashboard Will Never See

Between 50% and 80% of all social sharing happens off the record, through DMs, WhatsApp, Messenger, texts, and email forwards. When someone copies your link and texts it, it lands in Google Analytics as "direct traffic" with no trace of where it came from.

This rewires how you read a quiet post. Strong saves and sends with low public engagement might be your best content, spreading where you can't watch. Saves are the visible edge of an iceberg that's mostly underwater.

Real Social ROI Includes The Cost Almost Nobody Counts

Only 30% of marketers measure social ROI effectively, and 34% aren't sure they can at all. The formula is simple; the inputs are where people cheat:

ROI (%) = [(Revenue from Social - Cost of Social) / Cost of Social] x 100

For paid campaigns, ROAS = Campaign Revenue / Campaign Cost, as a ratio like 5:1. The trap in both is the cost side, because labor is the line nobody enters.

Ten hours a week at $50 an hour is $2,000 a month in labor before tools or ads. Count ad spend only and social looks miraculously profitable, which is how it fools people. Businesses do average $4 in revenue for every $1 spent on social, but that figure leans on brands with real attribution: UTM tags on every link, GA4 tracking referral traffic, consistent campaign naming. Build those into your content calendar and you're already ahead of most creators and small businesses.

Check Weekly, Not Daily, Because The Patterns Need Time To Form

Daily dashboard checking produces anxiety, not insight. The signals that matter emerge over weeks, and watching them by the hour just trains you to react to noise. A realistic cadence:

• Weekly (15 min): engagement rate trend, your top 3 posts and why, follower growth rate, reply rate.
• Monthly (30 min): format performance, real best posting times, audience shifts, referral traffic.
• Quarterly (1 hour): full ROI calculation, competitive benchmarks, strategy adjustments.

Building Autopilot taught me a version of this the hard way: confident, frequent output that nobody connects to an outcome is just a flat, dead post on a schedule. Measurement works the same way; checking constantly feels productive and changes nothing. Save daily checks for a live campaign or a crisis, and if you're saving hours with scheduling tools, don't spend them refreshing graphs.

FAQ

What counts as a good engagement rate?It depends on platform and formula. Roughly 1-5% is solid, but TikTok nano-influencers average 7.5% while Facebook brand pages sit at 0.15%. Only compare against benchmarks for your own platform and account size.

What is the difference between reach and impressions?Reach counts unique people; impressions count total views including repeats. If 100 people see your post and 20 see it twice, reach is 100 and impressions are 120. In 2026, Instagram replaced both with "Views."

Are saves better than likes for the algorithm?In 2026, yes, on Instagram and TikTok. A save signals utility, and utility gets rewarded with distribution, which is why reference content outperforms entertainment in reach per engagement.

How do I calculate my engagement rate?Use our free engagement rate calculator, or by hand: (Total Engagements / Total Followers) x 100, or divide by reach instead.

Stop Hoarding Data, Start Pulling Levers

The whole guide in five moves: pick one business goal, choose three to five metrics on the chain to it, set a 15-minute weekly review, reply to every comment for 90 days, reassess quarterly. Everything else is optimization you bolt on later.

If you're tired of jumping between native dashboards, a management tool centralizes everything. If you're an agency, you need reporting that clients who don't think in engagement formulas can still read. Sydium pulls analytics from Instagram, TikTok, LinkedIn, Facebook, X, and Threads into one dashboard, so you spend your time pulling levers, not hunting for numbers. Try it free and see what your data is telling you.