How to Build a Monthly Social Media Report Clients Actually Read

June 10, 2026 · 9 min read · by the Kadenzo team

Most social media reports are written to prove work happened. Twelve pages of screenshots, a metric for everything the API returns, a slide titled "Highlights" that's really just the three least embarrassing numbers. The client skims page one, says "looks good," and the agency wonders why the relationship feels fragile at renewal time.

The problem isn't effort. It's that an export is not a report. An export shows data; a report makes an argument — here's what we did, here's what it earned you, here's what we're doing next month and why. Clients renew arguments. Nobody renews a CSV.

This guide is the structure we'd use for any retainer client: five questions, answered in order, on as close to one page as you can manage. It assumes nothing about your tooling except that you can see your platform analytics and your website traffic.

A report is five answers, not forty metrics

Every monthly report worth sending answers these questions, in this order:

What were we trying to do? One short paragraph. The month's focus in plain words — "push the spring collection to existing followers and test whether LinkedIn can produce demo signups." If you can't write this paragraph, the report below it is decoration: there was nothing to measure against.
What happened? The numbers block — five to seven metrics, the same ones every month, each shown against last month and your trailing 90-day baseline. Numbers without history are trivia. The section below covers which metrics belong here.
Why did it happen? Two or three sentences of actual analysis per notable movement. "Engagement rate rose 0.8 points because the two carousel posts out-saved everything else we've published this quarter" is analysis. "Engagement was strong this month" is weather reporting.
What did it earn? Traffic, leads, signups, sales — whatever the client actually counts as the point. This is where most reports go vague, and it's the section the client's boss reads. It only works if your links were tagged before the month started; more on that below.
What's next — and what do you need from us? Next month's focus, one decision you want from the client, stated as a question they can answer in one email. Reports that end with a decision get replies; reports that end with a chart get silence.

That's the whole skeleton. Everything else in this guide is about making answers two and four cheap to produce and impossible to argue with.

The numbers block: fewer metrics, attached to their method

The fastest way to lose a client's trust is to have them ask "how is this calculated?" in a meeting and not have a crisp answer. The fix is mechanical: every rate in the report travels with its formula.

A defensible numbers block for an organic social retainer looks something like:

Posts published — by platform. Effort, stated plainly.
Reach (or impressions, per platform convention) — total and per-post median. The audience you actually earned.
Engagement rate — mean of per-post rates, calculated against reach, with the formula printed under the table. Our engagement rate calculator computes it this way and gives you a copy-paste methodology line precisely so the "how is this calculated" question takes four seconds to answer.
Saves + shares (or bookmarks + reposts, in each platform's own vocabulary) — the high-intent signals, separated from the likes that inflate totals but predict nothing.
Link clicks — from platform analytics, as the bridge into the traffic section.
Follower change — last, deliberately. It's context, not a result.

Three disciplines make this block trustworthy month after month:

Freeze the window. Calendar month or trailing 30 days — either works, but pick one and never switch. A window that moves is a number that can't be compared, and a client who notices the window moved is a client wondering what else moves.

Annotate outliers; don't erase them. A post went viral, or a post flopped because the platform throttled link posts that week. Show the average with and without the outlier, label it, and move on. Clients forgive variance. They do not forgive discovering, three months later, that "average" quietly excluded the bad weeks.

Pair every rate with an absolute. A rising engagement rate on collapsing reach is a shrinking account wearing a flattering metric. Showing 4.8% and the 41,000 reach it was earned on keeps the report honest — and keeps you from being surprised in the meeting.

The traffic block: won or lost before the month starts

Answer four — what did it earn — can't be reconstructed after the fact. If the links you posted in April weren't tagged, April's traffic is a blob labeled "social / referral" in analytics, and no amount of reporting effort will split it into "the spring campaign worked, the evergreen posts didn't."

The discipline is unglamorous:

Every outbound link gets UTM parameters, before publishing. One campaign name per initiative, shared by every channel; sources and mediums from a fixed vocabulary so they land in GA4's default channels where you expect. Our UTM builder generates the whole channel set at once and previews the GA4 classification per link — build the set when the campaign brief lands, paste the links into the content calendar, done.
Organic and paid get different mediums (social vs paid-social), never different sources. The report can then show "Instagram organic drove 312 sessions, Instagram ads drove 1,840" from one clean table.
Internal links never get tagged. Tagging your own navigation overwrites the visitor's real origin and quietly corrupts the exact table this section depends on.

With that in place, the traffic block is a four-line table — sessions, conversions, and conversion rate by channel, campaign by campaign — pulled straight from analytics with zero manual reconciliation. Without it, this section is guesswork wearing a table's clothes.

A worked example

Here's the spine of a real report, anonymized and trimmed. Boutique fitness studio, organic-led retainer, April:

Focus: April was the spring challenge launch — recruit sign-ups from the existing audience, with Instagram carrying the campaign and the newsletter as backup.

What happened: 14 posts published (9 Instagram, 3 LinkedIn, 2 newsletter sends). Instagram reach 41,200 (median 3,900/post). Engagement rate 4.8% by reach, mean of 9 posts — up from 3.9% in March; the three challenge carousels averaged 6.4% and out-saved everything else this quarter. (ER = (likes + comments + saves + shares) ÷ reach × 100, averaged per post.)

What it earned: 644 tracked sessions (spring-challenge-2026-04), 87 challenge sign-ups — 71 from Instagram organic, 16 from the newsletter. 13.5% session-to-signup.

Outlier note: April 22 Reel reached 18,400 (4.7× median) on a sound trend; shown excluded from the per-post averages above, included in totals.

Next + decision needed: Challenge retention content through May 15. We want to test one boosted post (€150) against the organic baseline — approve by Friday?

Six sentences of structure carrying every number that matters. The client can forward it to their boss without translating it — which is, quietly, the test a report has to pass.

What to leave out

Cutting is what makes the five answers readable. The usual suspects:

Follower count as a headline. It's the slowest-moving, least causal number you have. Keep it in the block as context; never lead with it.
Per-post screenshots. Link a content appendix for the client who wants to scroll; don't make every reader scroll past it.
Every metric the platform exports. Video average watch time, profile visits, taps-on-sticker — include a metric only when some decision would change if it moved. Otherwise it's padding between the client and the argument.
Comparisons across platforms. A 1.2% rate on X by impressions and a 4% rate on Instagram by reach aren't on the same scale — different formulas, different denominators. Compare each channel to its own history, and say so in the report.

The thirty-minute assembly

Run with discipline, the monthly report is an assembly job, not a writing job:

Pull the platform numbers into the standing table (10 minutes — the metrics and window never change, so neither does the query).
Pull the campaign table from analytics, already clean because the links were tagged on the way out (5 minutes).
Write the focus paragraph, the why-sentences, and the decision question (15 minutes — the only real writing in the document).

The hour-long version of this job is actually the cheap one. The expensive version is the day spent reverse-engineering untagged links and re-deriving last month's formula — every single month. Set the vocabulary once, freeze the window, tag before publishing, and the report becomes what it should have been all along: the easiest argument you make all month.