A monthly LinkedIn report template for clients that nobody has to decode

Print the band before you print the number

Give every metric a range before you give it a value. A client reading a bare number has no way to judge it, so they judge it against the only reference they have, which is a screenshot of somebody else's post or a benchmark from an article about a company nothing like theirs. The band removes that comparison by supplying a better one first.

Build the band from the client's own six months, not from an article

Use a trimmed range: take the last six monthly values, drop the highest and the lowest, and the band is the smallest and largest of the four that remain. Trimming matters because a single post that travels will drag a mean upward and set an expectation the client cannot repeat on purpose, and one quiet holiday month will drag it down and hide a real problem.

Here is the calculation with illustrative numbers you should replace with your own exports. Engagement rate over six months reads 2.9, 4.6, 3.4, 3.1, 6.8 and 3.8. Drop the 6.8, which was one post about pricing that reached far outside the usual audience, and drop the 2.9 from the month with a two week gap. The remaining four are 4.6, 3.4, 3.1 and 3.8, so the band is 3.1 to 4.6. This month came in at 4.0, which is in band, and the honest report says so in three words.

Illustrative band table using invented figures. The values are here to show the arithmetic and are not benchmarks for anybody.

Two rules keep the bands honest. Recalculate them every month on a rolling window rather than fixing them at the start of the engagement, and rebuild the whole band when a metric sits above it for three months running, because at that point the band is describing a client who no longer exists. A stale band produces a report full of good news that has quietly become ordinary.

The seven pages, in this order, every month

Seven pages, fixed order, same order forever. The order matters more than the content because a client who learns where things live stops reading the report and starts using it, and a report that changes shape every month gets read once and then skimmed for the graph.

The page order. Print it on the cover so the client knows what is coming.

Page three is the one most agencies skip and it is the one that saves the relationship in a flat month. Inputs are what you controlled and outcomes are what the platform decided, and a client who can see fourteen posts shipped, two ad sets rebuilt and a page section rewritten will read a flat outcome page very differently from a client shown only the flat outcome.

Page one is one sentence and one ask

Write the verdict as a sentence a busy person can read while standing up, then ask for exactly one decision. Most reports open with a dashboard and end with no request, which means the client receives information and gives back nothing, and the next month starts with the same ambiguity as the last.

August sat inside your own range on four of the five metrics. New followers came in below the band, and I think that is the two week posting gap rather than a reach problem, so I am not proposing a change on the strength of it. The one decision I need from you: whether to run the pricing objection post as an ad for thirty days at a fixed spend, with a stop rule if cost per enquiry passes the figure we agree today.Page one, worked example, invented client

Notice what that paragraph does. It gives the verdict before any number. It names a cause and labels the naming as an opinion rather than a finding. It declines to react to a single below band month, which is the discipline most retainers lack. And it ends with one decision carrying a stop condition, so the client is agreeing to something bounded rather than to an open ended experiment.

If page one ever needs to carry a return figure rather than an activity verdict, the numbers behind it need different handling and a stated confidence level on each one. The slide sequence for that conversation is set out in how to prove LinkedIn ROI to a board, and it is a different document from this report.

The block nobody includes: what this report does not claim

Write the limits of the report inside the report, as a short list, in the same typeface as everything else. This is the section missing from every published template, and it is the section that makes the other six pages credible. A client who has never been told what a report cannot show will assume it shows everything, and every gap then reads as either incompetence or concealment.

Read that block aloud on the first call of the engagement rather than letting the client discover it in month four. It costs three minutes and it converts every future awkward question into something you already answered. The credited enquiry line in particular needs the method behind it explained once, and the arbitration rules for it are set out in attributing pipeline to organic posts.

Separate what you shipped from what the platform decided

Split the report into inputs you controlled and outcomes you did not, on two separate pages, and never mix them in one table. Mixing them is what produces the two worst conversations in a retainer: taking credit for a month that went well by accident, and being blamed for a month that went badly for reasons nobody in the room controls.

• Inputs, page three: posts published and their formats, ads live and their spend, page sections rewritten, employee posts supported, hours worked against the retainer. All of these are countable and all of them are yours.
• Outcomes, page four: impressions, engagement rate, followers, sessions, enquiries. None of these are promised, all of them are reported against a band, and each one carries a one line explanation.
• The join between them, which is the useful part: name the specific input you believe moved the specific outcome, and label that belief with the word think. A named guess is useful and a hidden guess is not.

A useful test on the inputs page is whether a client could have counted it themselves from public information. If they could, it belongs on page three as a verifiable fact. If they could not, such as hours worked or drafts rejected, it still belongs there but it needs a line of context, because unverifiable inputs presented without explanation read as padding.

Clients running a company page alongside a founder profile will ask which one the report should cover. Report both, in the same band table, clearly separated, because the two behave differently and averaging them hides both. Whether a small company needs the page at all is a separate question covered in does a solo consultant need a company page.

The three positions a number can hold, and the move each one implies

Every metric lands in one of five states once you have a band, and each state has one correct response. Writing the responses down in advance stops the monthly meeting from becoming a negotiation about what the numbers mean, which is a conversation nobody wins and everybody dreads.

The response table. Agree it with the client in month one so nobody is improvising in month four.

The row that saves the most money is the first one. Retainers burn through tactics because a normal month with no stated band looks like a month that needs fixing, so something gets changed, and the change resets whatever learning was accumulating. A band gives you permission to do nothing, and permission to do nothing is worth more than most of the tactics it replaces.

How long the report should take, and what to automate first

Target forty minutes per client report and automate everything except page one. The arithmetic is straightforward and worth doing out loud, using assumptions you should replace with your own. Twelve client reports a month at ninety minutes each is eighteen hours a month and two hundred and sixteen hours a year. Getting each one to forty minutes is eight hours a month, which returns ten hours a month or a hundred and twenty hours a year to the business.

One caution on the tooling. A live dashboard link is not a substitute for the report, because a dashboard has no verdict, no band and no ask, which returns all the interpretation to the client. Send the dashboard as an appendix for the clients who want to poke at the raw numbers between meetings, and keep the seven pages as the thing you are actually paid to produce.