Most PR reports fail at the writing, not the measuring. The numbers are gathered, the template is opened — and then the document is assembled backwards: a confident summary first, evidence arranged underneath it, gaps quietly paved over. This guide covers how to write a PR report the other way round: records first, evidence laid out plainly, interpretation labelled as interpretation, the shortfalls written in the same tone as the successes, and the summary written last, when there’s something true to summarise.

Two companion guides hold the neighbouring ground, and this one deliberately doesn’t retread it. What a report must contain — above all, placements a reader can open and check — is set out in our guide to what a PR report should include. Reporting as a recurring discipline — report types, cadence, templates as fixed structure — is covered in our guide to PR reporting. Here, the subject is narrower and more practical: the act of writing, step by step.

Before You Write: Start from the Records, Not the Template

Gather the raw material before any sentence is drafted: the distribution report (if a service such as IMCWire carried the release, its report is the placement-list input — the list of where the release ran, with links), the release exactly as issued, your pitch log with dates and addressees, any coverage clippings, and the objectives document from the start of the period. Then reconcile. Open the placements. Check that the dates in the list match the period you’re reporting. Note any entry that doesn’t open, any duplicate, any item that turns out to be your own release re-posted rather than coverage. Discrepancies found now are footnotes; discrepancies found by your reader are credibility events. A report written on unreconciled records doesn’t contain errors of judgement — it contains errors of fact, which are worse, because they’re the kind a reader can prove.

Only after the records agree with themselves should the template come out. The template’s job is to receive verified material in a fixed order — it cannot verify anything itself, and beginning from it inverts the whole exercise.

The Steps, in Order

Step 1: Fix the Period and the Reader

Write down, literally, the period’s start and end dates and the name of the person (or the meeting) the report serves. Every inclusion decision later — does this placement count, does this activity belong — is answered by those two facts. A report without stated boundaries invites the quiet extension of the period to catch one late placement, which is how honest documents start their slide.

Step 2: Restate the Objectives as They Were Set

Quote the objectives from the document where they were agreed — not your memory of them, and never a version edited to match what happened. If the objectives were vague (“raise awareness”), reproduce the vagueness and note it; the report that admits its target was undefined teaches more than one that invents a precise target retroactively and hits it. If objectives changed mid-period, record the change and its date. The reader’s core question — did the work do what it set out to do? — can only be answered against the original wording.

Step 3: Lay Out the Placements Plainly

Now the evidence layer, written to the standard our contents guide sets out: every placement with its outlet, date, headline as published and a live link, in the fixed structure your reporting discipline uses — the same fields, in the same order, as last period’s report. Write this section with adjectives switched off. “The release ran in 14 outlets” is a sentence a reader can verify. “The release achieved extensive coverage” is a sentence a reader can only believe, and reports should deal in the first kind. Syndicated copies are labelled as syndicated copies; editorial coverage is labelled as editorial — the distinction is the report’s backbone.

Step 4: Separate What Happened from What You Think

Interpretation is not the enemy of an honest report — unlabelled interpretation is. Keep two visibly distinct layers: the record (what ran, where, when) and your reading of it (what it suggests, what you’d do next), headed or phrased so a reader always knows which one they’re in. Estimates live in the second layer and carry their method with them: “an estimated potential audience, summed from the outlets’ published figures” is honest in a way the bare number never is. The moment a guess puts on the uniform of a measurement, the document is lying even if the guess is good.

Step 5: Write What Did Not Happen

The passage most writers skip, and the one that makes the rest of the document believable: the target missed, the pitched feature declined, the outlet that never replied, the placement that ran with a fact wrong (and what you did about it). Write it in the same plain tone as the successes — no flinching, no burying it in a subordinate clause, no converting it into a lesson so fast the reader misses that it happened. A report with no failures in it is not evidence of a perfect period; it is evidence of an edited one, and experienced readers price that in.

Step 6: End with Next Actions That Follow from the Evidence

Each recommendation should be traceable to a line above it: the trade title that covered the story gets the next exclusive look; the angle that drew declines gets retired or rebuilt. If an action can’t point to its evidence, it isn’t a conclusion — it’s a preference, and it belongs in a strategy meeting, not in a report. Keep the list short. Three actions that follow from the record outweigh ten that follow from habit.

Step 7: Write the Summary Last

The summary goes at the front of the finished document and is written at the end of the process — assembly order and reading order are different things, and confusing them is the original sin of report writing. Written last, the summary is a compression: the period, the objectives, what ran, what didn’t, what happens next, in a handful of sentences that a reader who reads nothing else can rely on. Written first, it becomes a promise the evidence is then arranged to keep. If your summary and your evidence disagree, the summary is the one that’s wrong — go back and fix it, not the evidence.

Language Discipline: Verbs That Don’t Lie

Reports rarely lie in their numbers; they lie in their verbs. “Ran”, “was published”, “was carried”, “was opened”, “declined” — these describe events someone can check. “Drove”, “boosted”, “generated”, “delivered”, “leveraged” — these smuggle causation into the sentence, claiming the activity produced an outcome the report contains no evidence for. Traffic rose during the campaign month is an observation; the campaign drove traffic is a theory, and in a report it needs either evidence or a label saying it’s a theory. Do a dedicated pass through the draft hunting causal verbs and unearned adjectives — “significant” without a comparison, “strong” without a scale, “extensive” without a count — and for each one, either attach the evidence or downgrade the word to what the record supports. This pass, more than any formatting choice, is what separates a report from a brochure.

Presenting Figures Honestly

Every number in the report should arrive with its source and method attached, in the sentence or in a note: counted from what, over which dates, deduplicated how. A placement count from the distribution list, a referral figure from the site’s own analytics, an audience estimate from the outlets’ published figures — each labelled as what it is, because they are different species of number and adding them together produces fiction. Comparisons against the previous period are welcome and useful, but only while the definitions are unchanged; if what you count as a placement, or an item of coverage, has changed since last period, say so in the report and restate the earlier figure on the new basis or flag the break in the series. That discipline — same fields, same definitions, changes declared — is the cadence our PR reporting guide sets out, and it is what makes a run of reports worth more than any single one. On reach specifically: figures are estimates of potential audience, never counts of readers — the full treatment, including why advertising value equivalency is rejected, lives in the contents guide and isn’t repeated here.

A Short Fictional Example

The excerpt below is fictional — the company, its launch and all figures in it are invented to show the writing process, and are illustrative only. It shows three passages of one report, written the way the steps above describe.

Objectives, as set on 3 March: “Announce Tessellate’s invoicing update to accountancy-practice media; secure the announcement, accurately reported, in the trade titles read by practice managers.” (Note the absence of retro-fitting: the objective said nothing about national coverage, so the report won’t claim any was expected.)

What ran: The release was distributed on 10 March and ran in 14 outlets from the distribution list (links in the appendix); 11 carried the pricing fact correctly, one misstated it and corrected on request on 13 March. One trade title covered the update editorially, in its own words, on 12 March. (Facts, each checkable; syndication and editorial kept distinct.)

What did not happen: Neither national title pitched (two) took the story; both declined without giving a reason. The founder interview offered to three podcasts produced one recording, not yet published at the period’s end. (Written flatly, in the same type size as the wins.)

Next action, and its evidence: Offer the covering trade title first sight of the next update (evidence: Step “What ran”, line 3). The numbers inside this fiction are small on purpose: a real report’s credibility comes from exactly this plainness, at whatever scale the period actually produced.

Assembling and Checking: The Pre-Send Pass

• Open three links at random from your own placement list. The check buyers are told to run on a report works just as well on its author — before it goes out.
• Trace every figure back to the records in one step. If a number needs a memory or a spreadsheet hunt to justify, it either gets its source noted or comes out.
• Read the summary against the body. Every claim in it should be findable below, in stronger company.
• Check names, dates and titles as published — a report about accuracy cannot afford its own errors.
• Shorten. The reader is busy and the evidence is the value; most reports improve by losing a third of their commentary.

FAQs

Should the summary go first if it’s written last?

Yes — reading order and writing order are different things. The reader meets the summary first; the writer produces it last, as a compression of the finished evidence. The rule is about sequence of work, not position on the page.

What if there are no results to report?

Report the activity and the absence, plainly: what was sent, to whom, and that it produced no coverage in the period. A quiet period written honestly is information a decision-maker can use — the angle, the list or the timing needs to change. The same period inflated is a decision made on fiction, and the bill for it arrives in a later period, with interest.

The objectives were unrealistic. Can I quietly restate them?

No — that converts a planning failure into a reporting one. State the objectives as set, report against them honestly, and put your argument about their realism where it belongs: in the interpretation layer and the next-actions section, proposing better-set objectives for the next period. Readers forgive an unrealistic target; they don’t forgive discovering the target was moved.

How long should a PR report be?

As short as the evidence allows. The body should carry the period’s record and reading of it in a few pages; the clippings, full placement lists and raw data belong in appendices, where they support the report without swamping it. If a section exists because the template has a box for it rather than because the period produced something to say, write “none this period” in the box — that, too, is reporting.

Write the report you would trust if you were paying for the work it describes: records reconciled, evidence openable, facts and interpretation in separate clothes, the failures in the same tone as the wins, and a summary that the rest of the document can cash. That document doesn’t demonstrate success — it does something more valuable. It lets its reader judge.