Social media analytics: which metrics to track and how to report to leadership

A Sprout Social survey of 1,200 marketing leaders (2025) contains two numbers that explain half of our client conversations. 56% of leaders believe social media drives revenue. Only 44% think their team can measure it. The money keeps coming anyway: 80% are moving budget from other channels into social. In other words, leadership pays for what it believes and expects someone to prove it.

It can be proved. Just not with the metrics teams track out of habit.

What gets measured today

In the same Sprout Social report, teams define the return on social through engagement (68%), conversions (65%), revenue (57%), efficiency (55%) and discoverability (51%). The teams that leaders describe as expert lean on revenue and efficiency more than the rest do, and on engagement less.

That is the central gap. Engagement is the easiest thing to count, so everyone counts it. Revenue is hard to count, and the teams that do are more often the ones leaders call expert. According to Emplifi (405 marketers, report for 2027), 48% of teams can confidently show leadership the return on social with clear attribution. For the rest, the report rests on engagement.

A funnel instead of a list

A report that works is built as a funnel, not as a list of every number the platform exports. Reach: how many people saw it. Engagement, but only the kind that signals intent: saves, shares, direct messages. Visits to the site or the store. Inquiries and purchases. One or two numbers at each step, plus the conversion rate to the next one.

Why saves and shares specifically. The platforms themselves shifted weight from public reactions to private ones. In Socialinsider's 2026 benchmarks, comments per post fell 24% on TikTok and 16% on Instagram, while shares per post rose 45% and 12% respectively. Buffer, in its report on 52 million posts, writes that many platforms now prioritize private signals such as messages, shares and saves. Hootsuite's 2026 trends report calls follower count and engagement rate unreliable indicators outright. A like is politeness, a save is a plan to come back, and the funnel cares about the second one.

Tying social media to money

Perfect attribution in social does not exist and never will: someone sees a video, searches for the brand a week later and arrives two weeks after that on a friend's recommendation. Later's 2026 guide recommends a multi-touch model over last click. We agree, but for a small team that is a luxury. What works without complex systems:

• Promo codes and separate links for each channel and campaign. Influencer Marketing Hub (2026) found this is how most brands measure creator campaigns: 45.9% use promo codes, 26% affiliate links, 25% the platforms' built-in shops.
• A "how did you hear about us" field in the lead form with a "social media" option. Crude, but it gives you an order of magnitude.
• Branded search: if searches for the brand name rise after a campaign, that is a social media effect too, even if indirect.

Those three sources add up to one number worth bringing to leadership every month: cost per lead from social. It takes content and ad spend for the period and the number of leads tagged with promo codes, links and the "social media" answer in the form. The figure will be approximate, but it can be set against cost per lead from other channels, and that comparison is what leadership wants to see.

And one thing teams skip: according to Sprout Social, fewer than half of teams feed social data into a CRM. As long as a lead from social and a lead from the website live in different spreadsheets, proving revenue is impossible in principle. How we connect that data is described in our analytics and reporting service, and with social media management that link is part of the monthly report.

The review rhythm

HubSpot (2026) found that 44% of marketers evaluate campaigns weekly. We work like this: once a week we look at the funnel numbers and change nothing except for obvious failures. Once a month we compare with the previous month and decide what to double down on. Once a quarter we revisit goals and the platform mix. Judging a single post is pointless: the variance between individual posts is huge, and conclusions drawn from one lead to chasing random hits. How to set the goals the report is then built around is covered in social media strategy basics.

A report people read to the end

An executive reads the first page. It needs three things: what we promised at the start of the period, what happened in funnel numbers, and what we are changing. Everything else, the per-platform charts and the best posts, goes in an appendix.

There is a surprising gap here. In Sprout Social's research, 57% of marketing leaders believe executives trust their social team, while among executives themselves 74% say they do. There is more trust than marketers assume. At the same time, in the Sprout Social Index 2025, almost half of practitioners believe leadership only partly understands the value of social, and 41% say their biggest fear is not being allowed to publish what would perform best. A report built on a funnel and on promises helps narrow that gap. It moves the conversation from "they do not trust us" to "here is what we did with the money".

Which metrics to drop

Follower count as the headline number. A blended engagement rate that does not separate likes from saves. Reach with no link to the next funnel step. Any figure nobody can make a decision on. If a metric does not change a single action next month, it does not belong on page one.

How this funnel connects to paid campaigns is in the article on paid social, and the other tools that produce these numbers are collected in our overview of classic social media tools. If you would rather have us build the funnel, start on the for brands page.

Start simple: take your last report and cross out everything that does not lead to an inquiry or a purchase. What survives is your analytics. Put those numbers on page one of the next report and add a single line to them: cost per lead from social for the month.