Most "social media metrics" lists hand you twenty numbers and no way to tell what any of them means or whether yours is good. This is the reference that fixes both. It is a plain-English dictionary of the metrics a creator actually encounters in a native dashboard — reach, impressions and views, engagement rate, comments, shares, saves, watch time and completion, follower growth rate, brand mentions and sentiment, share of voice, response time, click-through rate, referral traffic, conversion rate, CPC, CPM, and ROAS — grouped the way the platforms group them and with the exact formula for each, so you can compute the ones your dashboard doesn't. But the part that makes a metric usable is the part these lists skip: a number alone is meaningless, and the single most misleading thing in social analytics is a benchmark, because the same metric reports wildly different figures depending only on the formula behind it and the vendor who published it. Buffer's own analysis of tens of millions of posts puts median engagement rates in the mid-single-digit percentages because it divides by reach; the big brand-account studies report tenths of a percent for the same platforms because they divide by followers — a ten-fold gap that is entirely definitional, not a difference in performance. So this guide spends as much time on how to read a metric as on what it is: which numbers are real signals and which are vanity, how to benchmark against your own trend line before any external figure, and how to confirm two numbers were computed the same way before you ever put them side by side. It closes on the quiet precondition no metric list mentions — that every one of these numbers is only as trustworthy as the consistency of the content underneath it, and a rate computed on three sporadic posts is noise no benchmark can rescue.
A social media metric is any single data point a platform reports about your content or account — a like, a view, a reach figure, a second of watch time, a click. That is the whole definition, and it is worth stating plainly because the word gets used as if it meant "something important," when most metrics are simply something counted. The skill is not collecting them; it is knowing what each one literally measures, how it is calculated, and whether the value in front of you is good — and that last question, whether a number is good, is the one almost every metrics list leaves you to guess at.
This guide is the reference that answers all three. It is the definitional, catalog-and-benchmark companion to its two neighbors and deliberately does not repeat them: the measurement process — the repeatable loop of setting goals, collecting data, and changing what you make next — lives in social media measurement, and the discipline of choosing three to five numbers per objective and setting targets on them lives in social media KPIs. This one is narrower and more practical: what each metric you will actually see in a dashboard means, the formula to compute it, and — the part that makes it usable — how to tell whether yours is any good without being fooled by a benchmark.
Before the catalog, the distinction that keeps it honest. A metric is any data point you can pull. A KPI — key performance indicator — is the narrow subset of metrics tied so directly to a goal that a change in the number would change what you do; every KPI is a metric, but most metrics are not KPIs for any given goal. A vanity metric is the trap in between: a number that feels like progress and rarely drives a decision. Raw follower count, total impressions, and cumulative likes are the classic three.
The fix for a vanity metric is not to ignore it but to pair it with a rate. Follower count on its own is a snapshot; follower growth rate is the signal. Impressions on their own only ever rise; click-through rate tells you whether anyone acted. Likes are the weakest engagement signal there is; engagement rate puts them in proportion to how many people actually saw the post. The operating test for whether a number belongs on your dashboard is one sentence — if it moved, would you do anything differently — and most of the time the raw count fails it while the rate passes.
Engagement metrics read whether the people who saw a post actually did something with it. The headline is engagement rate, most commonly (total engagements ÷ reach) × 100, where engagements are likes, comments, shares, and saves combined — the single number for "did this land." Within it, not all engagements are equal, and the modern reading weights them. Comments signal enough investment to type a reply, and here quality matters as much as count. Shares — measured as amplification rate, (shares ÷ followers) × 100 — are the strongest signal of all, because a share is an endorsement that carries your content to someone else's audience. Saves, read as save rate, (saves ÷ reach) × 100, are a close second: a save is a promise to come back, which the platforms increasingly treat as a high-value signal. The practical hierarchy for a creator is saves and shares first, comments next, likes last.
For video, two more engagement metrics have moved to the center as every platform reweights its algorithm toward retention. Average watch time is total watch time ÷ total views — how long, on average, people stayed. Completion rate is (completed views ÷ total views) × 100 — the share who watched to the end. These matter more each year because completion is increasingly the exact thing the algorithm optimizes for: a short with high reach and low completion is being shown widely and abandoned fast, and completion rate is your early warning that the hook or the pacing is losing people before the reach dries up.
Reach is the number of unique people who saw your content, each counted once. Impressions is the total number of times it was shown, repeat views included, so impressions always run at or above reach. The distinction is simple but constantly fudged: reach answers "how many people," impressions answers "how many times," and a high impression count with a flat reach means the same people are seeing you repeatedly rather than your audience growing. A complication worth flagging is that the major platforms are increasingly collapsing both into a single "views" number, and each platform defines a view on its own rules — so the label is drifting, and you should check what a given platform's "views" actually counts before comparing it to another's.
Follower growth rate — (new followers in a period ÷ followers at the start) × 100 — is the metric that should replace raw follower count in every report. It reads momentum rather than accumulated history: a large account that has stopped growing and a small one compounding fast look identical on a follower total and opposite on a growth rate, and the growth rate is the one that predicts where you will be in six months. Alongside it, audience demographics — the age, location, and interest breakdown the platform reports — is not a performance metric but a fit check: it tells you whether the audience you are growing is the one you set out to reach, which is the difference between growth that compounds into a business and growth that just inflates a number.
These read what happens off your own posts. Brand mentions is how often you are named across the platform, tagged or not, in a set period — the raw volume of the conversation about you. Brand sentiment, read as positive mentions ÷ total mentions, scores whether that conversation is favorable, hostile, or neutral, and a rising mention count with falling sentiment is a warning a volume metric alone would hide. Share of voice — your mentions ÷ (your mentions + competitors' mentions) × 100 — is the most strategic of the three, because it converts a raw count into a competitive position: it accounts for the fact that everyone's reach rises and falls with the platform, so it reads awareness relative to the field rather than in a vacuum. These are the metrics a listening tool exists to capture, and the full scorecard for them is in social listening metrics.
Social is a frontline support and reply channel now, and two metrics read it. Response time is the average time to first reply on a message, comment, or mention that needs an answer — total time to first response ÷ messages answered — and on social a slow number is visible to everyone watching the thread, not just the person waiting. Response rate is (messages replied to ÷ messages received) × 100 — the share of incoming conversation you actually engaged with. For a brand running support on social these are load-bearing; for a solo creator they matter less as formal metrics, but the underlying behavior does not: platforms consistently reward accounts that reply, and replying lifts the engagement on the original post measurably.
When the goal is action or money, the metrics get financial and, usefully, travel cleanly across platforms because you control the definition of the result. Click-through rate — (link clicks ÷ impressions) × 100 — reads whether the hook earned the click; it is not a conversion metric, a distinction people routinely blur. Social referral traffic is the volume of visitors social sent to your owned site, read in your web analytics via UTMs rather than in the platform. Conversion rate — (conversions ÷ clicks) × 100 — reads whether the destination and offer closed the loop, and splitting it from CTR is the whole point: a high CTR with a low conversion rate is a landing-page problem wearing a content problem's clothes. On the paid side, cost per click is spend ÷ clicks, cost per thousand impressions (CPM) is (spend ÷ impressions) × 1,000, and ROAS is revenue from ads ÷ ad spend — the cost-per-result metrics that belong at the top of any paid dashboard, above native counts that each platform measures on its own rules. Where attribution gets genuinely messy across channels, cross-platform campaign measurement is the companion read.
Here is the part that separates a usable metric from a decorative one, and it is the reason this page exists separately from the catalog any tool can give you. A metric in isolation — "2% engagement" — tells you nothing; it needs a reference point. But a benchmark is the single most misleading object in social analytics, because one headline figure hides two variables that completely swamp it: the formula behind the number and the sample it was drawn from.
The clearest illustration is engagement rate. Buffer's analysis of tens of millions of posts across hundreds of thousands of accounts reports median engagement rates in the mid-single-digit percentages — roughly 6% on LinkedIn, 5-6% on Facebook and Instagram, 4-5% on TikTok and Pinterest — because it divides engagements by reach. The large brand-account benchmark studies report tenths of a percent for the same platforms — well under 1% on Instagram, around 2-3% on TikTok — because they divide by follower count. That is close to a ten-fold gap on the same metric name for the same platforms, and almost none of it is a difference in performance; it is purely a difference in the denominator. A creator who reads a reach-based study and compares it to their own follower-based number will conclude they are failing when they may be doing fine, and vice versa.
Three rules fall out of this, and they apply to every metric, not just engagement. First, confirm the formula before you adopt a benchmark — a number computed by reach and a number computed by followers are different metrics that happen to share a name, and comparing them is meaningless. Second, compare like with like: the same platform's figure moves by a full point depending only on whether it was measured by followers, reach, or views, so a cross-platform "X beats Y" claim deserves scrutiny before you repeat it. Third, and most important, weight your own trend line above any external figure. Your last quarter, computed with your own consistent formula, is a more reliable benchmark than any industry average, because it is the one comparison where the formula and the sample are guaranteed to match. For the industry-specific version of this, see social media engagement rates by industry; for benchmarking against named rivals, social media competitor analysis.
The full catalog is for reference; the working set is much smaller. A solo creator growing an audience that feeds platform monetization should lead with four: follower growth rate, engagement rate (saves and shares weighted heavily), average watch time and completion for video, and the save rate on the formats they are trying to scale. Those are the numbers that both grow the account and satisfy the algorithm, and they are enough to run on. Conversion, CPC, ROAS, and the customer-service metrics mostly belong to brands running a funnel or a support channel; forcing them onto a creator dashboard adds noise, not insight. The honest shortlist beats the comprehensive one every time, because a dashboard that watches everything hides the two or three numbers that would actually change what you post next.
Kompozy is not an analytics tool, and it is worth saying so flatly: it does not compute your engagement rate, build your dashboard, or replace the native platform reporting and web analytics every metric above lives in — for dedicated tools, see the best social media analytics tools for creators. Reading the numbers is deliberately not its job. Kompozy addresses the thing one step earlier that makes the numbers worth reading at all, and this is the angle the measurement and KPI guides do not take: a metric is only as trustworthy as the consistency of the content underneath it.
Three of this page's own lessons point straight at that precondition. Benchmarking against your own trend line assumes you have a stable baseline — but a rate computed on three sporadic posts a month is noise, and a trend built from an erratic cadence measures how irregularly you shipped more than how your content performed. Comparing formats — the Reels-versus-carousels, short-versus-long question every creator eventually asks — is only valid if you produced enough of each format to make the comparison a real sample rather than a coin flip. And share of voice, which reads presence across the whole category, is structurally impossible to move if you are only consistently present on one platform. In each case the metric is being read against a foundation of content that most creators cannot sustain by hand.
That foundation is what Kompozy supplies. It is a full AI content generation and multi-platform publishing engine, not a repurposing tool: point it at a source — a long video, a URL, a topic — and it produces net-new content across the full range of output formats, from Persona Shorts and brand-exact HyperFrames carousels to Photo Posts, Quote Graphics, blogs, and newsletters, every output governed by one written Persona Brief so the voice holds steady and rendered on brand through HyperFrames so the look does not drift post to post. Then autopilot schedules and fans the approved queue across the eight social platforms plus blog and email behind a per-post review gate. The effect on your metrics is specific: enough volume that a rate is a real sample instead of a guess, enough consistency that a change in a number reflects the content rather than how erratically it shipped, and enough coverage across formats and platforms that format comparisons and share of voice become measurable at all. Measure with the platform tools and your CRM; use Kompozy to make the content behind the numbers trustworthy enough that the numbers are worth trusting.
A social media metric is just a data point; what makes it useful is knowing what it measures, how it is computed, and whether yours is good. Learn the small working set — reach and impressions for awareness, engagement rate with saves and shares weighted for resonance, watch time and completion for video, follower growth rate for the base, share of voice for position, and the conversion metrics for revenue — and compute the formula yourself so no tool's definition surprises you. Then treat every benchmark with suspicion until you know its formula and sample, and anchor on your own trend line above all of them. And remember the precondition underneath it all: a number is only as honest as the consistency of the content that produced it, which is the one thing no dashboard can fix after the fact.
There is no universal set — the ones that matter depend on your goal. But the metrics worth knowing how to read are: reach and impressions (awareness), engagement rate plus saves and shares (resonance), average watch time and completion rate (video), follower growth rate (audience), brand mentions, sentiment and share of voice (brand health), response time (service), and click-through rate, conversion rate and ROAS (action and revenue). Track three to five tied to your current goal and treat the rest as context. For a creator, engagement rate, saves, watch time and follower growth rate usually lead.
Reach is the number of unique people who saw your content, counting each person once. Impressions is the total number of times it was displayed, including repeat views by the same person — so impressions can exceed reach. Many platforms now report a combined "views" number instead, and each platform counts a view on its own rules, so the labels are drifting. The practical rule: reach answers "how many people," impressions answers "how many times," and you should never treat one as a substitute for the other.
The common formula is (total engagements ÷ reach) × 100, where engagements are likes, comments, shares and saves. But there are two widely used denominators that answer different questions: dividing by reach reads how compelling a specific post was to the people who saw it, while dividing by follower count reads how active your existing audience is over time. The two produce very different numbers, so pick one, define it once, and never mix them in a comparison.
Because they measure the same thing differently. A benchmark is only a number until you know the formula and the sample behind it. Buffer's analysis of tens of millions of posts reports median engagement rates in the mid-single digits because it divides engagements by reach; large brand-account studies report tenths of a percent for the same platforms because they divide by followers. That gap is definitional, not a performance difference. Always confirm which formula a benchmark used, compare like with like, and weight your own trend line above any external figure.
Vanity metrics are numbers that feel like progress but rarely change what you do — raw follower count, total impressions, and cumulative likes. They are not useless; they are context worth a glance for direction. The fix is to pair each with a rate that carries meaning: follower count becomes follower growth rate, impressions become click-through rate, likes become engagement rate. A raw count tells you how big; the rate tells you whether what you are doing is working.
Kompozy is not an analytics tool — it does not compute these metrics or build your dashboard; the native platform reporting and your web analytics do that. Kompozy sits one step earlier, on the thing that makes the metrics trustworthy in the first place: consistent, on-brand, high-volume output. A rate computed on three sporadic posts is noise no benchmark can fix, and a format comparison is meaningless unless you produced enough of each format to compare. Kompozy is a full AI content generation and multi-platform publishing engine that supplies that volume and consistency, so the numbers you read reflect your content rather than how erratically it shipped.
Social media metrics are the individual data points platforms report on your content: reach, impressions, engagement rate, comments, shares, saves, watch time and completion, follower growth rate, brand mentions, sentiment, share of voice, response time, click-through rate, referral traffic, conversion rate, CPC, CPM and ROAS. Each reads one thing — reach counts unique viewers, engagement rate reads resonance. The catch is formula: the same metric swings ten-fold depending on the denominator and the vendor, so benchmark against your own trend line first and a formula-matched external figure second, never a raw number alone.
Get started → · ← All guides · Compare Kompozy vs other tools