Social Media Analytics for Business

Strong social media analytics answers three questions on demand. First: what is working right now? This means identifying the top posts, formats, and themes from the last 7 to 30 days, with enough context (impressions, engagement rate, click-through rate) to repeat the win. Second: what is changing? Trends matter more than snapshots. A 4% engagement rate is meaningless until you know last quarter was 6%.

Third: how do we compare? A post that drew 90 likes reads differently once you know the median tracked competitor account drew 13.1 likes per Facebook post across 54 accounts in June 2026. Without a competitive benchmark, internal numbers are abstract. This is why most mature programs pair their own analytics with competitor tracking, and why the engagement per post benchmark page publishes the counts with their sample sizes attached. Knowing your number is table stakes. Knowing your relative number is strategy.

Every major platform ships native social media analytics: Meta Business Suite for Facebook and Instagram, the X Analytics dashboard for Twitter, LinkedIn Page Analytics, and TikTok Analytics. These are free, accurate, and granular for your own accounts. They are also siloed, hard to compare across platforms, and almost useless for tracking competitors.

Third-party tools fall into three buckets. Publishing-led suites (Buffer, Hootsuite, Sprout Social) bundle scheduling with analytics. Listening platforms (Brandwatch, Talkwalker, Meltwater) focus on mentions and sentiment at scale. Competitive intelligence tools (including Oppira) focus on tracking what other brands are doing across Facebook, Instagram, and Twitter/X without needing API access to their accounts. Most teams need a mix. The native tools tell you what you did. Third-party tools tell you what everyone else did.

Inside Meta Business Suite, Page Insights gives you reach, impressions, engagement, follows, and content performance per post. Useful, but most teams drown in it. Three views matter more than the rest. The Content tab lets you sort posts by reach, reactions, comments, shares, and link clicks over a custom range. Sort by shares, not reactions. Shares are the strongest organic distribution signal Facebook offers, and 1.49 shares per post across 54 tracked competitor accounts is the June 2026 line to clear.

The Audience tab shows demographics, geography, and follower growth. Watch net follower change weekly. A flat follower count with rising reach usually means the algorithm is favoring your content. A falling follower count with rising reach means people are unfollowing after seeing you in feed (a quality warning). The Benchmarking tab compares your page against up to 100 similar pages you select. It is rough, but free.

Ads Manager is where paid performance lives. The metrics that matter for most B2B campaigns are CPM, CTR (link), CPC (link), landing page view rate, and cost per result (lead, signup, demo). For B2C, add ROAS and frequency. Frequency above 3.0 in a single week usually indicates creative fatigue and rising CPMs.

The Facebook Ad Library is the most underused free tool in social media analytics. It shows every active ad from every page, including ad copy, creative, format, and (in some regions) spend ranges and reach. You can study a competitor's full active ad set in 10 minutes. The limitation: you do not see performance, only existence. Oppira fills part of this gap by tracking which competitor posts (organic and boosted) are getting outsized engagement, so you can infer what is working without guessing from creative alone.

Volume in the library needs a reference point before it means anything. Tracked advertisers ran 9.37 Meta creatives a month across 75 advertisers in June 2026, so a competitor holding thirty live variants is running several times the typical programme. With your own house in order, the next platform is where most modern social engagement actually happens.

Inside Instagram Insights, Reels expose plays, reach, likes, comments, shares, saves, average watch time, and total watch time. Two metrics predict distribution: average watch time as a percentage of total length, and shares plus saves combined. A Reel with 65% average watch time and a save rate above 1% will almost always get a second algorithmic push 24 to 48 hours after posting.

Comments and likes matter less than they used to. The Instagram algorithm appears to weight saves and shares as stronger personal signals because they imply intent to revisit or recommend. When auditing Reels, sort by reach first, then look at watch-through rate within the top 10. The best-performing Reels almost always share a hook in the first 1.5 seconds, a clear payoff, and a length under 30 seconds.

Feed posts (single image, carousel, photo) still matter, but for different reasons. Carousels generate the highest save rate of any organic format on Instagram, often 2x to 4x a single image. Stories drive replies and DMs, which Instagram rewards as high-intent signals. Watch tap-forward rate per slide. If slide 3 of a 5-slide story sees 40% tap-forward, that is a hook problem in the first two slides.

Hashtag analytics inside Instagram Insights now show reach from hashtags as a single bucket, not per-tag. To get per-tag performance, you need third-party tools or manual experiments. A practical method: post the same content style with two different hashtag sets across 8 weeks, then compare reach-from-hashtags averages. Most accounts find that 3 to 5 well-chosen tags outperform 30 generic ones, contradicting older Instagram playbooks.

Native Instagram Insights only shows your own account, which is the single largest blind spot in most social media analytics setups. To understand competitors, you need to track their public posts over time: what formats they use, posting frequency, top-performing themes, and engagement trends. Doing this manually for 5 competitors burns 4 to 6 hours a week. Tools like Oppira track this automatically and surface the outlier posts (a Reel that is suddenly pulling 10x normal engagement, for example) so you can investigate why. With Instagram covered, the third pillar of most B2B social programs is Twitter/X, which behaves differently from both.

Inside the X Analytics dashboard (analytics.x.com), the core metrics per tweet are impressions, engagements, engagement rate, link clicks, profile visits, follows, and bookmarks. Bookmarks have become a strong signal in 2026 because, like Instagram saves, they indicate intent to revisit. A tweet with 50,000 impressions and 800 bookmarks is doing different work than a tweet with 50,000 impressions and 800 likes.

Engagement rate on Twitter/X is genuinely lower than on other platforms. Median rates around 0.03% to 0.09% are common for accounts under 100,000 followers. Do not benchmark against Instagram. A 1% engagement rate on X usually means a tweet went viral. Profile visits per tweet is the leading indicator that someone may follow or click your bio link, which is where most B2B Twitter conversion actually happens.

Threads (multi-tweet posts) get separate impression counts per tweet in the chain. Watch the drop-off from tweet 1 to tweet 2. A healthy thread retains 40% to 60% of impressions on the second tweet. Below 25% means your hook tweet promised something the thread did not deliver, or the thread was too long.

Replies are an underrated growth surface on X. Replying to larger accounts in your space often generates more profile visits than your own tweets, especially under 10,000 followers. There is no native dashboard for reply performance, so you have to track it manually or with a third-party tool. The metric to watch is profile visits driven from replies, not likes on the replies themselves.

X audience insights are weaker than Meta's. You see follower count, follower growth, and basic demographic estimates. For richer audience analysis (interests, overlap with competitor audiences, follower quality), you need third-party tools. The most useful exercise: pull your follower list and the follower lists of 3 to 5 competitors, calculate overlap, and identify who follows competitors but not you. Those accounts are your warmest expansion targets. Across all three platforms, the next question becomes: which numbers should drive the business?

Awareness stage: impressions and reach. These tell you how many people saw your content. Use reach (unique accounts) for organic strategy and impressions for paid frequency management. Track share of voice (your brand mentions vs total category mentions) at this stage too.

Consideration stage: engagement rate, average watch time, save rate, and profile visits. These tell you whether the audience cared enough to interact. Engagement rate alone is noisy. Pair it with watch time on video and saves on static content for a fuller picture.

Action stage: link clicks, click-through rate, landing page conversion rate, and cost per lead (for paid). This is where most teams lose the thread, because clicks attributed to social are often undercounted by 30% or more due to dark social (private shares, DMs, screenshots). We will return to this in the ROI section.

If you can only watch one metric per platform week to week, choose the leading indicator that predicts everything else. On Facebook, it is shares per post, and 1.49 shares per post across 54 tracked competitor accounts is the measured June 2026 line. On Instagram, it is saves plus shares per Reel. On Twitter/X, it is profile visits per tweet. On LinkedIn (briefly, since it is not the focus here), it is dwell time per post.

These four metrics share a property: they all measure intent beyond passive consumption. A like is cheap. A share, save, profile visit, or long dwell costs the user something (social capital, attention, a click). That cost is what the algorithms reward. Track these and most other metrics will follow. With KPIs framed, the next question is the one that gets miscalculated more often than any other in social.

Engagement rate by reach: (engagements / reach) x 100. This is the most accurate for organic content because it measures interaction among people who actually saw the post. Use this when reach data is available (Instagram, Facebook, LinkedIn).

Engagement rate by impressions: (engagements / impressions) x 100. Slightly lower than ER by reach because impressions count repeated views. Use this when comparing paid campaigns or when reach is not exposed (Twitter/X often only gives impressions).

Engagement rate by followers: (engagements / followers) x 100. The most generous formula and the most commonly quoted in agency reports. It is useful for benchmarking competitors when you cannot see their reach, but it overstates performance because reach is almost always smaller than follower count. Engagement rate by post: average ER per post over a period, weighted equally regardless of reach. Useful for content audits, misleading for performance summaries.

The cleanest way to compare platforms is to convert everything to rates and indexed values. Engagement rate by impressions works on every platform. Cost per click works on every paid platform. Click-through rate from social to your site is the same metric everywhere because the destination (your site) is the same.

Avoid summing absolute metrics across platforms in headline numbers. A report that says 'we generated 4.2 million impressions across social' is not wrong, but it is not useful either, because a Twitter impression and a TikTok impression are not the same product. Better to report per-platform numbers side by side, then a single rolled-up business metric (clicks to site, conversions, pipeline sourced) that is calculated identically regardless of source.

Interaction counts are the exception worth knowing about, because a like is a like on every platform even when the scale differs several-fold. Instagram measured a median 58.6 likes per post across 51 tracked competitor accounts, Facebook 13.1 across 54, and LinkedIn 19.2 across 33 in June 2026, which is a comparison you can make without normalising anything.

There are four common approaches. First, native dashboards plus a manual spreadsheet. Free, accurate, and slow. Most teams burn 6 to 10 hours a month on this. Second, a publishing tool with built-in analytics (Buffer, Hootsuite, Sprout Social). Good for your own accounts, weak for competitor data. Third, a BI tool (Looker, Power BI) connected to platform APIs. Powerful, expensive in setup time, and requires API access you may not have for competitor accounts.

Fourth, a competitive analytics tool that pulls public competitor data automatically alongside your own. Oppira falls in this bucket: it tracks competitor posts and engagement across Facebook, Instagram, and Twitter/X without needing API access to those competitors, and reports them in a unified view. The right answer for most teams is a combination: native tools for owned accounts, a competitive tool for competitor and category data. Once data is unified, the harder skill is knowing which numbers to ignore.

Follower count is the classic vanity metric. A page can grow followers via giveaways or buying them, with zero downstream business effect. But follower count becomes actionable when paired with engagement rate over time. A flat or falling engagement rate on a rising follower count is a clear quality warning.

Impressions and reach are similarly slippery. Reaching a million people is meaningless if none of them are buyers. Reach becomes actionable when filtered by audience match (right geography, right industry, right job titles for B2B) and paired with click-through rate. Likes are almost always vanity in 2026. Saves, shares, and comments contain real signal. Likes are pattern-matching.

The measured counts make that hierarchy concrete. The median tracked account drew 58.6 likes and 0.8 comments per Instagram post across 51 accounts in June 2026, so likes arrive by the dozen and the typical account goes several posts between comments. The scarce signal is the one worth reporting.

A simple filter for any metric in a report: ask 'so what?' three times. Engagement rate went up 12%. So what? More people are interacting with our content. So what? The recent product-demo Reels are pulling 2x the saves of brand posts. So what? We should publish two product-demo Reels per week instead of one and pause the lifestyle content. That is an actionable metric chain.

If a metric cannot survive three rounds of 'so what?', it does not belong in your weekly report. It belongs in an annual review or in nowhere at all. Trim ruthlessly. A four-metric weekly report that drives decisions beats a 30-metric report that drives meetings. With the noise filtered, the next question is the one executives always ask.

First, last-click attribution via UTM parameters, the campaign tags Google Analytics documents and reads. Easy to set up, badly biased against social. It works for direct-response campaigns (paid social driving a checkout) and breaks for brand-led demand. Use it as a floor, not a ceiling.

Second, multi-touch attribution. Tools like HubSpot, Salesforce, and dedicated MTA platforms assign fractional credit across touchpoints. Better than last-click, still blind to dark social. Most B2B teams find that multi-touch raises social's credited contribution by 1.5x to 3x compared to last-click.

Third, self-reported attribution. Add 'How did you hear about us?' to demo-request and signup forms. This catches dark social better than any tracking-based method. The numbers will not tie out perfectly with your analytics, and that is fine. The goal is directional truth, not GAAP-accurate accounting. Most B2B teams that add this question discover social is 2x to 4x larger as a source than their UTM data suggested.

The basic ROI formula is (revenue attributed - cost) / cost. The honest version pairs three numbers: tracked revenue (from UTMs and platform conversion APIs), self-reported revenue (from form questions), and pipeline influence (deals where social was a touchpoint, even if not first or last).

Costs should include paid spend, tooling, agency or freelance fees, and a conservative estimate of internal headcount time. Most teams forget headcount, which makes their ROI look better than it is. A social team of 2 FTEs at $90,000 fully loaded each is $180,000 a year. That number belongs in the denominator. With ROI framed honestly, the final piece is making the whole system run on a clock.

The weekly review is for the social team and the immediate manager. Audience: 2 to 5 people. Time: 30 minutes. Metrics: top 5 posts of the week (with brief notes on why), bottom 2 posts (with hypotheses), one leading indicator per platform, and any competitor outliers worth investigating. Output: a one-page document and a list of 2 to 3 changes for next week's calendar.

The weekly cadence is where competitor tracking pays off most. If a competitor's Instagram Reel suddenly hits 10x their normal engagement, you want to know within 48 hours so you can adapt or counter while the topic is hot. Manual checking will not catch this. Tools with automated alerts (Oppira's AI-powered alerts, for example) flag these outliers as they happen. The weekly meeting is where you decide what to do with that information.

The monthly review is for marketing leadership. Audience: 5 to 15 people. Time: 45 to 60 minutes. Metrics: full KPI scorecard against targets, format and theme breakdown, paid versus organic split, share of voice trend, top 3 wins and top 3 losses with diagnoses, and changes for next month. Output: a 5 to 8 page deck or doc.

This is where format decisions get made. If carousels are outperforming Reels on saves but Reels are winning on reach, leadership needs to see the tradeoff and pick a direction. Video was 41.8% of Instagram posts across 51 tracked competitor accounts in June 2026, which is a useful check on any proposal to move the whole calendar to one format. Monthly is also where you spot creative fatigue, channel saturation, and competitive shifts that take longer than a week to materialize.

The quarterly review is for executives and cross-functional leaders. Audience: 10 to 30 people. Time: 60 to 90 minutes. Metrics: pipeline and revenue attributed to social (tracked plus self-reported), customer acquisition cost from social channels, share of voice versus top 3 competitors, audience growth quality, and a strategy update for the next quarter.

Quarterly reports answer the question: should we spend more, less, or the same on social, and on which platforms? They should make the case with numbers a CFO can audit. Benchmark figures belong here too, quoted with their sample size and their as-of date, so a claim about the category can be checked rather than trusted. They should also make space for what the numbers do not capture: brand sentiment shifts, competitor moves, platform changes (algorithm updates, new ad formats, regulatory shifts) that will shape the next 90 days. A program with these three cadences running cleanly will out-decide a program drowning in dashboards every time.