A client report answers "what did we achieve and what did it cost", which is a different question from the weekly update's "what is happening". Lead with the answer, not the data: what happened, what it means, what you recommend, and only then the numbers that support it. Send it to whoever approved the budget, not just your day-to-day contact, because that person decides renewal and is usually the one who never sees your work. The most common failure is not a bad report - it is a report that takes four hours to build, degrades by month three, and stops entirely by month six.
Reporting is the part of agency work most likely to be done badly by people who are good at everything else.
The pattern is consistent. Month one produces a beautiful report. Month two is nearly as good. By month four it is a rushed export with a paragraph on top, and by month six the client is asking where it got to. Nobody decided to stop; the reporting simply lost, every month, to work that had a deadline attached.
That matters more than it looks, because the report is frequently the only artefact the budget holder ever sees. They did not attend the calls, they did not read the updates, and their sense of whether you are worth renewing is formed almost entirely by a document you built in a hurry.
This guide covers what a report is for, the structure that works, how to choose metrics, how to report a bad month, and how to make reporting cheap enough that it survives a busy quarter.
The reader you are actually writing for
Not your day-to-day contact. The person who approved the budget, has never spoken to you, and will decide whether to renew.
Reports and updates are different things
Worth separating, because conflating them produces documents that do neither job.
An update answers "what is happening?" Short, frequent, operational, aimed at your day-to-day contact. Covered in our guide to the client communication plan.
A report answers "what did we achieve, and what did it cost?" Longer, periodic, tied to objectives and money, aimed at whoever approved the spend.
The audiences are different people with different questions. A marketing manager wants to know the campaign shipped. Their director wants to know whether the money produced anything. Sending both the same document means one of them is reading something that was not written for them.
Two documents, two jobs
The weekly update keeps your contact informed. The monthly report keeps the budget holder convinced. Sending one to both means one of them is reading something written for someone else.
Who the report is actually for
The single most useful reframe: write for the person who will decide whether to renew.
That is usually not your day-to-day contact. It is their manager, or the finance lead, or a founder - someone who has never spoken to you, has forty seconds of attention, and whose entire impression of your value comes from this document.
Three consequences follow.
No unexplained jargon. Not because they are unsophisticated, but because they do not live in this discipline daily. Expand acronyms, and translate metrics into consequences.
Money must appear. A report with activity and no reference to spend, budget, or return leaves the reader to do the arithmetic, and they will do it uncharitably.
The first paragraph carries everything. Assume that is all that gets read, because frequently it is.
Who reads it, in practice
Three people, with different needs. Your day-to-day contact wants confirmation that things are on track. Their manager wants to know the money produced something. Whoever approves the budget wants one sentence they can repeat to someone else.
A report written for the first reader alone is the most common mistake, and it is why so many good agencies lose accounts they were servicing well - the person deciding never saw anything that spoke to them.
The structure
Six sections. The order matters more than the contents.
1. The answer, in three sentences
Open with the conclusion. What happened, what it means, what you recommend.
Demo requests grew 18% this month against a target of 15%, driven mainly by the two new landing pages. Cost per enquiry fell from £61 to £48. We recommend shifting next month's budget from display to search, where the cost per enquiry is roughly half.
A reader who stops here has everything they need to feel good about the spend and knows what you are asking for. Everything below is evidence.
The instinct is to build up to the conclusion. Resist it - this is a business document, not an argument.
2. Performance against what you agreed
Not every metric you can produce. The specific ones you committed to.
A simple table with the metric, the target, the actual, and the direction of travel. Three to five rows. If you agreed a target at kickoff and it is not in this table, the report has quietly moved the goalposts, and sophisticated readers notice.
3. What we did
Deliverables, briefly. This is the section agencies over-invest in, because it is the part that feels like proof of effort.
It is not proof of value. Keep it to a scannable list and let the results section do the work.
4. What we learned
The section that separates a report from a receipt.
What did the month teach you about their business, their audience, their funnel? Even in a quiet month there is usually something: a page that underperformed, a segment that behaved unexpectedly, an assumption that turned out wrong.
This is where you demonstrate you are thinking rather than executing, and it is the most common differentiator between an agency that gets renewed and one that gets replaced by a cheaper supplier doing the same activities.
5. What we recommend
One to three specific recommendations, each with a reason and a rough size.
Be directive. "We could consider testing the checkout flow" invites nothing. "We recommend testing the checkout flow next month - it is where 40% of drop-off happens, and it is roughly two weeks of work" is a decision the reader can make.
6. Time and budget
For a retainer: hours used against the allocation, and what is queued. For a project: budget consumed against the plan.
Agencies leave this out to avoid drawing attention to cost. That is exactly backwards. A client who can see the hours understands what the fee buys; a client who cannot is left guessing, and people guess unfavourably. It is also the mechanism that makes overage a conversation rather than a surprise.
The three-sentence test
Before sending any report, read the first three sentences alone and ask whether someone who read nothing else would know what happened, whether it was good, and what you want them to do.
If not, the report is organised around your activity rather than their decision - which is the most common structural fault in agency reporting and the easiest to fix.
Choosing what to report
The failure mode is reporting everything you can measure, which buries the two numbers that matter and makes the reader do the analysis.
Three tests for including a metric:
Does it connect to something they care about commercially? Impressions rarely do. Qualified enquiries do.
Would you change what you do based on it? If a number moving would not alter your plan, it is decoration.
Did you agree it at the start? Metrics introduced mid-engagement, especially after a bad month, read as goalpost-moving even when they are not.
Swydo's guidance on client reporting makes the related point that a report should be built around the client's objectives rather than around what the tools export easily - which is the actual reason most reports are bloated. The dashboard produces forty metrics, so forty metrics appear.
Vanity metrics deserve a specific warning. Reporting impressions, reach or sessions when conversions are flat is a choice the reader will eventually notice, and when they do, every previous report becomes retrospectively suspect. If the good numbers are the shallow ones, say so plainly: "traffic is up 40%, conversion is flat, and the traffic is not the problem."
What a good report is worth
An agency that reports well is not doing better work than one that reports badly. It is making its work visible to the person who decides whether to keep paying for it - which, in a business where the buyer rarely sees the delivery, is most of what renewal turns on.
The uncomfortable version: agencies lose accounts they were servicing well, because the budget holder never saw evidence of it. The report is frequently the only artefact that person encounters all year.
Reporting a bad month
The test of a reporting process is what it does when results are poor.
Lead with it anyway. The bad number goes in the first three sentences. A reader who discovers it on page three concludes it was hidden, which is far more damaging than the number.
Distinguish what you control from what you do not. A market shift, a seasonal dip and a mistake are different things and should be described differently. Do this factually - the moment it reads as excuse-building, it stops working.
Bring a diagnosis, not just a result. "Conversions fell 12%" is a fact. "Conversions fell 12%, and it tracks almost exactly to the checkout change on the 8th - we are reverting it this week" is a report.
Say what would change your mind. "If this does not recover by the end of the month, we would recommend pausing the campaign and reallocating." Volunteering the exit condition is the strongest possible signal that you are optimising for their outcome rather than your invoice.
Our guide to difficult client conversations covers the same structure applied verbally, and the report should never be the first time a client hears bad news - anything material gets a call before the document lands.
The failure nobody plans for
Month one produces a beautiful report. By month six it has stopped. Nobody decided to stop - reporting simply lost, every month, to work with a deadline attached.
That is a design problem rather than a discipline problem, and the fix is to make the report cheap enough that a busy month cannot kill it.
Making it cheap enough to survive
Everything above is achievable in month one. The question is month nine, and this is where reporting actually fails.
Automate the collection, write the interpretation. The numbers should assemble themselves. The three sentences at the top cannot be automated and should not be - they are the entire value. If your report takes three hours, roughly two and a half are being spent on assembly, which is the wrong half.
Fix the template. Same six sections, same order, every month. Variation costs time and makes month-on-month comparison harder for the reader.
Keep it to two pages. A longer report is not more thorough, it is less likely to be read. Detail belongs in an appendix or a live dashboard for the one client in ten who wants it.
Make status continuously visible so the report is not the only window. A client who can see progress and hours in a portal at any time is not depending on the monthly document for reassurance, which lowers its emotional stakes considerably - and makes a genuinely quiet month easy to report honestly.
Put it in the calendar with an owner. Same date each month, one named person. Reporting fails because it is nobody's deadline.
Never let the report deliver bad news first
Anything material - a missed target, an overrun, a problem - gets a call before the document lands. A report should confirm what the client already knows, never surprise them.
The quarterly business review
The monthly report handles operations. The quarterly review is where renewals are actually decided, and agencies that only ever meet their day-to-day contact are one reorganisation away from losing the account.
Get the budget holder in the room. The whole point. Someone who has never spoken to you is deciding whether to keep spending, and a document is a weak substitute for a conversation.
Change the altitude. Not last quarter's activity - the direction of travel. What has been learned about their business, what is working, what should change, and what you would do with more or less budget.
Bring a recommendation with a number attached. A QBR that ends without a decision has been a status update with better catering. One specific proposal, sized and priced, gives the meeting a purpose.
Ask what has changed on their side. New priorities, new people, new pressures. This is frequently where the next engagement comes from, and it is the question most likely to reveal that the thing you have been optimising stopped mattering two months ago.
Sixty minutes, quarterly, for any account above a meaningful revenue threshold. It is the single highest-return meeting in an agency's calendar and it is routinely skipped because nobody asked for it.
Reporting by engagement type
The six-section structure holds; the emphasis shifts.
Retainers. The dominant question is "what am I getting for this", so hours against allocation and work delivered carry the most weight. A retainer report that omits the burn figure invites the client to assume the worst about value. This is also the report most likely to determine renewal, because retainers renew by default until someone decides otherwise.
Performance work - campaigns, SEO, conversion. Results and recommendations dominate. Include the baseline every time, not just the current figure; a reader three months in has forgotten where you started, and "conversion is 2.1%" means nothing without "up from 1.2% in March".
Project work. Reports are milestone-shaped rather than monthly. The useful additions are budget consumed against plan and any change orders raised, so the final invoice is never a surprise.
Discovery or advisory. The deliverable is thinking, so the report is largely the "what we learned" section. Resist padding it with activity - a short report full of insight reads better than a long one full of meetings attended.
Automating the assembly
The reason reporting degrades is that it competes monthly with billable work and has no external deadline. The fix is to reduce the effort rather than to increase the discipline.
The numbers should assemble themselves. Whatever produces your metrics - analytics, time tracking, project data - should feed the report without anyone copying figures between systems. Manual assembly is where the four hours go, and it is the half that adds no value.
The template should be fixed. Same six sections, same order, every month. Redesigning it each time is invisible effort that nobody asked for.
Only the interpretation should be written fresh. The three sentences at the top, the "what we learned" section, the recommendation. That is perhaps twenty minutes of genuine thinking, and it is the entire value of the document.
Set a standing date and an owner. Reporting fails because it is nobody's deadline. The same person, the same working day each month.
An agency that gets report production down to forty minutes will still be reporting in month twelve. One where it takes four hours will not, regardless of intention - and month twelve is exactly when the report matters, because that is when someone is deciding whether to renew.
Getting the report read
A report nobody opens has the same value as no report.
Put the three-sentence summary in the email body, not only in the attachment. Many readers never open the file, and those three sentences are the ones that matter most.
Send it the same day every month. Predictability is what turns it into something expected rather than something that interrupts.
Name the file usefully. Northwind-Report-2026-08.pdf, not Report_final_v3.pdf. It will be filed, forwarded and looked for later.
Put one thing in it that requires a response. A recommendation with a decision attached. A report that asks nothing invites no engagement, and engagement is what tells you whether it is landing.
Follow up once on silence. Not to chase praise - to check the recommendation landed. "Did the search reallocation make sense? Happy to talk it through" often surfaces that the report was never read by the person who needed to read it, which is worth knowing.
The one change worth making first
If your reporting is currently a monthly scramble, change one thing: put the answer in the first three sentences.
Not the data, not the activity, not a build-up. What happened, what it means, what you recommend. Everything else in this guide is refinement; that single change is most of the difference between a report that gets read and one that gets filed.
Cadence
Monthly suits most retainers. Frequent enough to catch drift, infrequent enough that there is something to say.
Quarterly business reviews for larger accounts, in addition. Longer, strategic, with the budget holder present - this is where renewals are really decided, and an agency that only ever meets its day-to-day contact is one reorganisation away from losing the account.
Weekly reporting is almost always a mistake. It becomes a status update in a report's clothing, costs four times as much to produce, and trains the client to evaluate you on a timescale where normal variance looks like failure.
What good looks like
A two-page document, arriving the same day each month, that opens with three sentences a busy director can act on, shows performance against agreed targets, says what you learned, recommends something specific, and shows what the hours went on.
It takes forty minutes to produce because the numbers assemble themselves. It goes to the budget holder as well as your contact. And in a bad month it says so in the first paragraph.
That is not a sophisticated reporting practice. It is a repeatable one, which is the only kind that is still running in month twelve - and month twelve is when it matters, because that is when someone is deciding whether to renew.