What actually belongs in a monthly client report

There is a particular kind of agency report that runs to forty slides, takes two days to build, and is opened once for ninety seconds. Everyone involved knows this. It gets made anyway, because a thin report feels like a thin month, and because nobody wants to be the agency that sent one page.

The irony is that the forty slide report is usually the one that loses the account. It buries the answer, it never says what went wrong, and it consumes the two days that could have been spent on the work it is reporting.

One number per objective

Start by writing down what the client hired you to change. Usually there are two or three things: leads, awareness in a specific market, share of voice against a named competitor. Then give each one a single number, this month against last, and a sentence saying why it moved.

That page is the report. Everything after it is evidence, and evidence is only read when somebody disputes the answer.

If you cannot name the objective, that is the finding, and it is worth more to the relationship than another chart. A client who cannot tell you what success looks like is a client who will eventually decide, privately and without warning, that you did not deliver it.

Cut anything you would not act on

A useful test for every chart: if this number were twice as high, or half as high, what would we do differently next month? If the honest answer is nothing, the chart is decoration.

Impressions usually fail this test. So does follower count, most months. So does any metric that only moves when the platform changes how it counts. Keeping them in the deck trains the client to skim, which means the numbers that matter get skimmed too.

• Keep: the numbers tied to the objective, the work delivered, the spend, and what you are doing next month.
• Cut: platform metrics nobody chose, screenshots of things the client has already seen, and any slide whose title is just a month.
• Move to an appendix: the detail one person on the client side wants and nobody else opens. They will thank you and everyone else will stop scrolling past it.

Say what did not work

A report with no bad news reads as a report that is not being written honestly, and clients notice sooner than agencies expect. Not immediately, but around month four, when the numbers are always green and the business has not obviously changed.

Naming the thing that underperformed, and what you changed because of it, buys more trust than another green arrow. It also gives you somewhere to put the months when a campaign genuinely does not land, which will happen, and which is far easier to discuss if it is not the first time you have ever raised a problem.

The month you first report a failure well is usually the month the client starts treating you as an advisor rather than a supplier.

Write the summary for the person who was not in the meetings

Your day to day contact already knows what happened. The report is read by their manager, and sometimes by a finance director deciding whether to renew. Write the first paragraph for that person: what the objective was, what moved, what you are doing next. No acronyms, no platform jargon, no reference to conversations they were not part of.

Generate it, do not assemble it

The reason reports are late is almost never the writing. It is the gathering: screenshots of coverage, exports from three platforms, the spend figure from someone in finance, all pulled together by hand in the last two days of the month by whoever has the least leverage to refuse.

If the work is already recorded as it happens, the coverage logged the day it lands, the spend attached to the campaign, the posts sitting on the calendar they were scheduled from, then the report is a review step rather than a production job. That is the whole argument for keeping the work and the reporting in the same system, and it is worth roughly two days a month per client.

Send it before the meeting

A report read live is a report nobody has thought about. You spend the call narrating slides, the client reacts to the first number they see, and the useful conversation never happens.

Send it two days early with a one line summary in the email body, and the meeting becomes a discussion about what to do next, which is the part you are actually paid for. It also gives a client who is unhappy somewhere private to be unhappy first, which is almost always better than discovering it in front of six people.