growth & reporting · 28 august 2026 · 7 min read

what should a monthly marketing report include?

Ask what should a monthly marketing report include and you'll usually be handed a template: sessions, users, followers, impressions, click-through rate, a line chart per channel, a summary slide. Twelve pages. Every number present, no question answered.

The test of a report isn't whether it's thorough. It's whether, having read it, you know what to do differently next month. Most monthly reports fail that test — not because the numbers are wrong, but because nobody decided what the report was for before they started filling it in.

Worth saying plainly: we produce monthly reports for clients, so we have an obvious commercial interest in them looking valuable. Most of what follows argues for a shorter report than the one you're probably being sent, and one whole section is a list of things your analytics genuinely cannot tell you. Weight our opinion accordingly.

a report exists to support one decision

A monthly marketing report has a single job: to make next month's decision easier than it would have been without it. That decision is almost always one of three — keep going, change something specific, or stop spending here.

That framing kills most of the standard template on contact. Impressions don't help you choose. Follower count doesn't help you choose. A month-on-month bar chart of sessions, with no sense of which sessions were the kind you wanted, doesn't help you choose either. If a number can go up or down without changing anything you'd do, it belongs in an appendix, or nowhere.

the three questions a monthly marketing report should answer

Everything worth including maps to one of these. If a section doesn't, cut it.

1. Did more of the right people arrive? Not traffic — qualified traffic. Which channels sent it, and whether that mix moved. A drop in total sessions alongside a rise in enquiries is a good month, and a template that leads with sessions will report it as a bad one.
2. Did any of them become an enquiry? The count of real conversion events — calls, forms, bookings, replies — with the path they came in on. This is the number the business actually runs on, and it's usually the smallest number in the document.
3. What changes next month, and why? A short list of specific actions tied to what rows one and two showed. Not "continue to optimise engagement" — a named thing to start, stop or test, with the evidence beside it.

Three questions is the whole structure. A page each is plenty. The pressure to pad comes from a reasonable fear — that a short report looks like light work — but a long report that hides its own conclusion is worse than a short one that states it.

what to actually put in it, section by section

A version we'd defend, in order:

• The decision, first. Two or three sentences at the top: what happened, what it means, what we're changing. Written so someone can read only this and be correctly informed.
• Enquiries, with sources. The count, the change on last month, and where they came from. Include the ones that arrived by phone or in person if you can capture them — those are invisible to analytics and are frequently the majority.
• Traffic and its quality. Volume by channel, plus one quality signal you've agreed on in advance — pages per session, time on the money pages, whatever fits.
• What we published, and how it did. The month's output listed against its performance. This is where you find out whether one topic outperformed the rest by an order of magnitude, which is the most useful thing a content report ever tells you.
• Spend and cost per enquiry, if you're running ads. One figure, tracked over time, beats a dashboard of ad metrics — and it's the number that decides whether the budget you set is still the right one.
• What we couldn't measure. The section almost nobody writes. It's next.

the section nobody includes: what the report can't know

Every monthly report is built on tools with documented limits. Most reports present their numbers as though those limits don't exist. Four worth naming in writing, because each one changes how much weight a figure deserves.

1. Your history has a shelf life. On a standard Google Analytics 4 property, the retention period for event data can be set to 2 months or 14 months — the longer options are 360 only. Two details matter for reporting. Increasing the setting "is applied to data that you have already collected and that you have not already deleted", so a year spent on the shorter setting is gone for good, not recoverable by changing it later. But the limit is narrower than it's often described: Google states the setting "does not affect standard aggregated reports" and "only affects explorations and funnel reports". Your headline monthly numbers survive; the flexible digging-around does not.

2. Small numbers get withheld deliberately. Google applies data thresholds so that individual users can't be identified from demographics, interests or other signals — rows are suppressed when counts are low, and a data quality indicator tells you it happened. Small business reports are exactly where this bites, because the segments are small by definition. An empty demographic breakdown often means "too few people to show", not "nobody came".

3. The visitors who declined cookies may simply be absent. Google can model the behaviour of users who decline analytics cookies, but behavioural modelling has published prerequisites: the property must collect at least 1,000 events a day with consent denied for at least 7 days, and have at least 1,000 daily users sending events with consent granted on 7 of the previous 28 days — and Google notes that meeting them still doesn't guarantee eligibility. Our own reading of those thresholds, not Google's statement about small business: most small sites will never come close to 1,000 daily consenting users, so for them the modelling never switches on and the declining visitors are missing from the report rather than estimated within it. If you run a consent banner, your traffic is undercounted by an unknown amount, and no honest report claims otherwise.

4. Attribution is a setting, not a fact. In GA4, "the first click, linear, time decay, and position-based attribution models are no longer available as of November 2023", and — the part that catches people out — "changing the reporting attribution model applies to historical and future data". Change the model and last month's report changes too. Meanwhile Meta lets you choose an attribution model — and, for standard attribution, the attribution settings — at the ad set level, under its own rules.

Which is why your ad platform and your analytics will never agree, and why chasing the discrepancy is wasted time. They are two systems counting under two different sets of rules. Pick one as the number of record, say in the report which one you picked, and stop reconciling.

the honest bit about month one

Your first monthly report will not tell you much, and a good one admits it. Organic content is a three-to-six month compounding play; a single month of it is noise with a chart on top. The early reports are for confirming the machine runs — things went out on time, they were indexed, people arrived — not for judging returns.

The report earns its keep from about month three, when there's enough history to separate a trend from a fluctuation. That's also when the question changes from "is it working" to "which part is working", which is the question that actually pays. We've written separately on how to tell whether marketing is working at all — the reporting habit here is what makes that judgement possible rather than a guess.

so what should a monthly marketing report include?

The decision on page one. Enquiries with their sources. Traffic and one agreed quality signal. What you published and how it did. Cost per enquiry if you're paying for traffic. A named list of what changes next month. And a short, unflattering paragraph on what the tools couldn't see this month.

That's five or six pages, and it's more useful than the twelve. The measure of a report is not how much it contains — it's whether someone who reads only the first page ends up making the same decision as someone who reads all of it. If those two people disagree, the report is badly built, however complete it looks.

If you're being sent something longer that never quite says what to do, you're not being under-served on data. You're being under-served on conclusions.