Reporting guide
How to build a monthly LinkedIn report your client actually reads
Most of these reports get skimmed for ninety seconds and then forwarded. Build for that and the whole thing gets shorter.
Build for ninety seconds
Assume your client opens the report on a phone, between two meetings, and reads until the first thing that requires effort. Then assume they forward it to somebody who has no context at all.
That gives you two constraints, and they decide everything else. It has to make sense without you in the room, and the most important thing has to be in the first screen. Which rules out the two most common openings: a title slide, and a summary of what you did. What you did is not the report. What happened is the report.
What belongs on the first screen
Four numbers, each with its change since the last report, and nothing else. The change is the part that carries meaning. 12,400 followers is a fact and 12,400 followers with 780 more than last month is a report.
Then one line of plain language saying whether the month was good, flat or bad. Write it as a sentence, not a colour. A client who has to interpret a traffic-light graphic will interpret it wrong in whichever direction they were already leaning.
Why reach alone does not work as the headline number
Impressions move with the algorithm, with the time of year and with whether one post happened to travel. A month with one lucky post and three weak ones outperforms a month with four solid ones on impressions alone, and reporting that as progress means next month reads as a decline you then have to explain. Engagement rate over the period is the more stable headline, with impressions underneath it as context.
The section most reports leave out
Almost every client report says what happened. Very few say what you concluded from it and what you are changing because of it.
That section is short. Two or three lines, each naming a specific thing you will do differently, and each attached to something in the numbers above. The carousel format outperformed text posts three months running, so November runs two of them. The posts published before 8am consistently underperform, so publishing moves. This is also the section that makes the report worth paying for, because it is the only part a dashboard cannot produce. Anyone can send numbers.
And then check it next month
Open the following report by going back through those two or three decisions and saying whether they happened and what came out of it. A recommendation that gets quietly dropped three months in a row was either a bad recommendation or one you never actually explained, and a report that revisits its own conclusions surfaces that instead of repeating it.
What to cut
Most reports are long because cutting requires a decision. These four go.
Screenshots of individual posts
They take up a screen each and say less than a line of text about why the post worked.
Follower demographics
Interesting exactly once, at the start of an engagement, and never again on a monthly cadence.
Anything you cannot act on
If a number would not change what you do next month, it is context, and context belongs in an appendix nobody opens.
Your own activity log
Nine drafts written and four calls held is an invoice line, not a report line. Keep it out and the report reads as being about the client.
What this costs you to produce by hand
Assembling one of these from screenshots and platform exports runs twenty to thirty minutes per client once you know what you are doing. At fifteen clients that is five to seven hours a month, roughly seventy hours a year, and it lands in the last week of the month when the team is already full.
The part that takes the time is not the writing. It is collecting the same numbers from the same places in the same period for every client, one at a time. Which is worth knowing, because that part is the part that can stop being manual, and the analysis section is the part that cannot.
The outline, in order
Copy this structure and fill it. One page, or one screen and a bit. Sections two to four are two lines each. If any of them runs to a paragraph, it is doing analysis that belongs in section five.
Getting the numbers without the collecting
Sections one and four are the mechanical parts, and that is where the seventy hours sit. whoranks tracks the four headline numbers per profile on one scale, in the same period, with the change since last month already calculated, and lets you put the people your client gets compared to next to them. Sections five and six stay yours. That is the part clients are paying for.