How to report bad results to clients without losing them
Every agency has had the Sunday night where the numbers are in, they are down, and nobody wants to write the email. So the report goes out late, softened, padded with a chart that happens to be green. Three weeks later the client is "just having a chat" with another agency.
Learning how to report bad results to clients is not a soft skill. It is a retention skill. Clients rarely leave because one month underperformed. They leave because it underperformed and they found out late, vaguely, or from their own bank statement. Delivery is the difference between a difficult conversation and a lost retainer, and it starts with how often you report at all.
This guide covers what to say, in what order, in what format, and how to build a reporting rhythm that makes bad months survivable.
Last updated: August 2026
Key takeaways
- Report a bad month in the cycle it happened in, not the next one. Clients forgive results. They do not forgive surprises.
- Lead with the number, then the cause, then the action. Burying a decline under three good metrics reads as spin.
- Declines are normal in real reporting data. Across 59 headline period-over-period comparisons sitting in ReportsMate's own report cache, 22 were down, 18 were up and 19 were unchanged. More than a third of headline numbers move backwards in any given period.
- A bad month inside a consistent rhythm is a conversation. A bad month after three months of silence is a resignation letter.
- Delivery format matters as much as wording. A decline explained in an email the client opens beats a decline sitting in a dashboard nobody logs into.
What's in this guide
- Why do clients really leave after a bad month?
- How bad is a bad month, really?
- What is the right way to report bad results to clients?
- How should you structure the report itself?
- What do you actually write in the email?
- What makes delivering bad news to clients worse?
- How does reporting cadence protect you before results go bad?
- Frequently asked questions
Why do clients really leave after a bad month?
Clients leave over communication, not performance. A down month with a clear explanation and a named next step is a normal part of a working relationship. A down month the client discovers themselves, weeks later, tells them you either did not notice or did not want to say. That is a trust problem, and it is the one that gets you fired.
Your client is not comparing your results to a benchmark. They are comparing their confidence in you this month against their confidence last month, and every silent week costs you a little of it. Client retention difficult conversations are usually won or lost before the conversation starts.
We built ReportsMate email-first for exactly this reason. After years around agency reporting, the pattern was consistent: the polished dashboard the client had access to almost never got logged into, so the only report that landed was the one in the inbox. When results dip, a report nobody opens is worse than none at all, because you believe you communicated and the client knows you didn't.
Two terms worth being precise about here. Reporting cadence is the fixed rhythm at which reports go out (daily, weekly or monthly) regardless of whether the news is good. Retainer churn is the rate at which those recurring contracts cancel. Cadence is one of the few churn inputs an agency fully controls, and it costs nothing to fix. If you want to size what a single cancellation is actually worth to you, run the numbers in our churn cost calculator.
How bad is a bad month, really?
Most reports contain at least one metric going the wrong way, and that is ordinary. We looked at the headline period-over-period comparisons stored in ReportsMate's report cache: of 59 comparisons across live client reports, 22 were negative, 18 positive and 19 flat. Declines slightly outnumbered gains. That is what marketing data looks like at a period level, and it is worth saying out loud, because agencies routinely treat normal fluctuation as a crisis and then transmit that panic to the client.
Some of what looks like a bad month is not performance at all:
- Attribution and modelling changes. Conversion counts in Google Analytics 4 shift with attribution settings, data thresholding and consent behaviour, all documented in Google Analytics Help. The campaign did not change. The counting did.
- Auction pressure. A competitor entering the auction moves your impression share, CPC and cost per conversion without you touching a setting, as Google Ads Help documents.
- The learning phase. Meta ad sets re-enter the learning phase after significant edits, and delivery is less stable while they do. That is set out in the Meta Business Help Centre.
- Search volatility. Ranking and click declines often trace to a broader search update rather than anything on the site, which Google Search Central documents and publishes.
None of that is an excuse to hand the client. It is context, and the difference matters: an excuse explains why the result is not your fault, context explains what actually moved so you can both decide what to do next. Clients accept context. They can smell an excuse.
What is the right way to report bad results to clients?
Report the decline first, explain the cause second, and commit to a specific action third. Do not warm them up. Do not open with the three metrics that happened to rise. Six steps, in order:
- Say the number in the first line. "Leads were down 22% in August, 41 versus 53 in July." No adjectives.
- Say what caused it, or say you are still finding out. "Still confirming" is a legitimate answer on day one. Inventing a cause you have not verified is not.
- Separate what changed from what you changed. Auction, seasonality and tracking sit in one bucket, your own decisions in another. Own the second one plainly if it applies.
- Give the trend, not just the point. One month against one month is noise. One month against a rolling three-month average is information.
- Name the action, the owner and the date. "We are rebuilding the two worst-performing ad groups this week and you will see the effect in the 12 September report."
- Say what you need from them. Budget, creative, a landing page fix, faster lead follow-up. A bad month is the one moment a client is most willing to act.
That order answers the client's real question, which is never "what happened" but "do you have this under control". Reporting a bad month in marketing is a demonstration of control, whether you are presenting PPC results or explaining an organic dip.
How should you structure the report itself?
A bad-month report needs a different shape from a good-month report. Same data, different emphasis: less celebration, more diagnosis and forward plan. Here is the structure that holds up in a difficult month.
| Report section | What goes in it | What it does for the client |
|---|---|---|
| Headline summary | The metric that matters most, current versus prior period, stated plainly | Removes any suspicion you are hiding something |
| Context | Seasonality, auction shifts, tracking or attribution changes | Separates signal from noise |
| What we changed | Every optimisation made in the period, with dates | Proves the retainer was worked, not idled |
| What worked | The parts that did perform, kept short and honest | Stops the report reading as a total loss |
| Diagnosis | Your actual read on why the number moved | The paragraph clients read twice |
| Next 30 days | Specific actions, owners and dates | Converts anxiety into a plan |
| What we need from you | Budget, assets, approvals, faster lead handling | Makes the client a participant, not a spectator |
Two structural rules. The headline number goes above the fold, before any chart. And the metric set stays identical to every other month: quietly swapping the featured metric from leads to impressions in the month leads fell is the most obvious tell in agency reporting, and experienced clients know it on sight. Deciding which metrics belong in the report once, then never changing them, protects you here.
What do you actually write in the email?
Write it short, factual and structured, with the decline in the subject-adjacent first sentence. Long apologetic emails read as guilt. Here is a workable shape, edit to your own voice:
Hi Sarah,
August was down. Leads came in at 41 against 53 in July, a 22% drop.
Two things drove it. Our top campaign lost impression share to a new competitor bidding on your core terms, and the mid-month site migration briefly broke conversion tracking, so some of the gap is measurement rather than lost leads. Both are quantified below.
Cost per lead held at $68, so the account did not get less efficient. It got smaller.
This week we are restructuring the two affected ad groups and re-verifying tracking end to end. You will see the first read on that in the 12 September report.
One thing we need from you: the updated pricing sheet for the landing page, ideally by Friday.
Full report attached below.
Notice what is missing. No "unfortunately", no "as you may be aware", no throat-clearing before the number. Honest client reporting means the first sentence carries the news, because that is the sentence everyone actually reads. Notice too that the email is the report, not a cover note pointing at a login screen.
What makes delivering bad news to clients worse?
Most damage in a bad month comes from the delivery, not the data. The recurring mistakes:
- Delaying. Late bad news is read as concealed bad news.
- Padding. Opening with vanity metrics that rose while the metric that pays their invoices fell.
- Changing the goalposts. Introducing a new "primary KPI" in the month the old one dropped.
- Blaming the platform for everything. Some of it usually is the platform. All of it never is.
- Over-apologising. Three paragraphs of contrition invites the client to conclude something is genuinely wrong.
- Going quiet afterwards. The follow-up report matters more than the bad one. Skipping it confirms the worst reading.
- Sending a dashboard link instead of an explanation. A link asks the client to diagnose their own bad month, and they will diagnose it uncharitably.
That last one matters most. A login-required dashboard is a fine analysis tool and a poor communication tool, and the gap shows up precisely when the news is bad. It is why we argue email reports beat dashboards for client-facing communication. That is our product's vantage point, so judge it against your own open rates.
How does reporting cadence protect you before results go bad?
The reporting rhythm you set in a good month determines how a bad month lands. If your client has heard from you weekly for six months, a down month is one data point inside an established relationship. If they hear from you every eleventh week, the down month is the relationship.
Frequency is the first lever. Weekly or fortnightly reporting surfaces a decline while it is still a small number, which lets you raise it as an observation rather than a confession. Monthly-only reporting concentrates all your credibility into twelve moments a year, and one of them will be bad.
Format is the second. In ReportsMate's own delivery data, of 2,463 client report emails confirmed delivered, 759 registered an open, roughly three in ten. Treat that as a floor rather than a ceiling, since open tracking relies on a pixel and undercounts. The point is not the number, it is that an emailed report tells you whether the client ever saw it. A dashboard login you do not monitor tells you nothing, and in a bad month you need to know.
Two things are worth having in place before you need them. White-labelling means the report carries your agency's branding rather than the tool's, and sender identity means it arrives from your own domain: bad news should never turn up from software the client has never heard of. Consistent platform coverage matters just as much. Google Analytics 4 is the most-connected platform across ReportsMate accounts, live on 30 of the 43 clients with an active connection, with Meta Ads, Google Ads, Search Console and Google Business Profile behind it. Reporting every connected channel every period, including the ones having a bad month, is what makes the good months credible.
To set that rhythm without adding hours to your week, see how it works or compare plans and pricing. Automation will not write the difficult email for you. It makes sure the difficult email is never also a late one.
Frequently asked questions
Q: How do you tell a client their campaign is not working?
A: State the result in the first sentence, give your honest read on the cause, and name the specific change you are making with a date attached. Softening language before the number only delays the information the client came for. If you do not know the cause yet, say so and give a date by which you will. Clients accept "I am still diagnosing this, I will have an answer Thursday" far more readily than a confident explanation that turns out to be wrong next month. What you must never do is skip the reporting cycle while you work out what to say.
Q: Should you report a bad month before the client asks?
A: Always, and inside the normal reporting cycle rather than as a special announcement. Volunteering bad results is the single strongest trust signal an agency has, because it proves your reporting is not filtered for flattery. Waiting to be asked converts a performance issue into a credibility issue, and credibility is the harder one to rebuild. A fixed reporting schedule does this work for you, since the report goes out on its date whether the news is good or bad and nobody has to decide whether to send it.
Q: How often do marketing metrics actually go down?
A: Frequently enough that a single down metric should not trigger a crisis response. In ReportsMate's report cache, 22 of 59 headline period-over-period comparisons were negative, 18 positive and 19 unchanged. Comparing one period against a rolling average rather than the single prior period is the fastest way to tell a genuine trend from ordinary variance. This also gives you a better answer for the client, because "down against last month but level against the three-month average" is a materially different message from "down".
Q: What if the bad results are the agency's fault?
A: Say so directly, early, and pair it with the correction and a date. "We over-weighted budget to a campaign that stopped converting, we caught it in week three, here is what we have moved" preserves far more of the relationship than a vague reference to market conditions. Clients are more forgiving of admitted, corrected errors than most agency owners assume. What they do not forgive is discovering the error themselves, which is a different and much worse conversation.
Q: How do you stop bad results from becoming client churn?
A: Report consistently before the bad month arrives, deliver the news in a format the client actually opens, and follow up on the stated fix in the very next report. Most churn traces to communication gaps rather than single bad periods, so the fix is structural, not conversational. Working out what one cancellation costs your agency in annual revenue with a client lifetime value calculator is usually enough to justify the reporting rhythm on its own.
Final tips on reporting bad results to clients
Knowing how to report bad results to clients comes down to three habits. Report on a fixed rhythm so nobody has to decide whether to send. Lead with the number, the cause and the action, in that order, every time. Follow up on what you promised in the very next report, because that is the one that decides whether the client believes you.
The agencies that survive bad months are not the ones with better results. They are the ones whose clients never had to wonder what was going on. Build the reporting system while things are going well, because you will not want to build it the week you need it.
Stop losing your Sundays to client reports. Start your free 14-day trial - no credit card, no setup fees, cancel anytime. Your clients get branded, white-labelled reports in their inbox automatically, in the good months and the bad ones.