Hub · Learn  /  Updated June 2026

How to explain an off week to a client without losing the account

An off week in an ad account is a communication problem layered on a diagnosis problem. The communication problem is the one that kills retainers. Before the call, verify the dip was real and identify which layer moved: the cost to reach people, the share who click, or the share who buy after clicking. On the call, lead with what moved and what it means in the client's terms, not with what you plan to test next. Propose one named next step tied to that specific layer. After the call, send a two-paragraph written note. Clients do not leave over one bad week. They leave when bad weeks arrive with vague answers.

Six-step order for the off-week client call: check measurement first, name which layer moved, state variance or real break, give one named next step, set a report-back date, send the written note within 24 hours.

Was the week actually off?

Before you open the call, answer this question for yourself. Not roughly. Specifically.

Pull your account data for the week in question and for the four weeks before it. Look at the three numbers that tell the full story: CPM (what you paid to show the ad to a thousand people), CTR (the percentage of people who saw the ad and clicked), and CVR (the percentage of people who clicked and then completed the desired action, whether a purchase, a form fill, or a booked call). The sibling article on why ads stop working walks through how to read each of these layers in sequence. The short version: a CPM spike means something changed in the auction. A CTR drop means the message stopped connecting. A CVR drop means the offer or post-click path has a problem.

Now ask whether the dip falls inside your normal week-to-week variance or outside it. Most accounts have weeks that run 10 to 20 percent above or below their rolling average without anything actually changing in the account. Seasonality, day-of-week mix, a single large order shipping attribution into a different window. None of these are breaks. They are noise. If the week lands inside your normal band, you do not have an off week. You have a normal week that looked bad because the client did not have the comparison data.

If it lands outside the band, then a layer actually moved. Name which one before the call. Not a guess. The specific metric, the magnitude of the change, and whether the shift is holding into the current week or recovering. One week of wobble is variance. A level shift that persists into week two is a break worth naming.

Do this check before you pick up the phone. An operator who walks into a client call with the specific number that moved, and the approximate reason it moved, lands completely differently from one who says "yeah the week was rough, we're going to look into it."

What do I say instead of "we'll test more creative"?

Stop using that sentence. Not because it is wrong, but because it trains the client to expect a new guess every time the numbers dip. Repeat it enough and the client stops believing you understand what is happening in the account. They start wondering whether you are managing the campaigns or just reacting to them.

"We'll test more creative" is a non-answer wearing the costume of a plan. It says: I do not know what moved, and I am going to try something and see. Clients hear that. They may not say so on the call, but the next Friday Slack comes with a shorter leash.

Here is what the two answers sound like back to back.

Guessing answer
"The week was soft. We're going to test some new creative angles and see if we can get it moving again."
Named answer
"Click-through rate dropped 30 percent but CPM held flat. The auction is fine. The message stopped connecting. Two new angles go live Thursday. Written update by Friday."

Both sentences end in the same action: new creative. The named answer does not take longer to say. It takes thirty seconds to say. The difference is that it shows you looked at the numbers and know what moved. The client hears that you know exactly which layer moved and why new creative is the right response to that specific layer. They are not hearing "we'll try something." They are hearing a diagnosis followed by a prescription.

Practice giving the named answer even when the answer is simple. The account will thank you more than the client does.

How do I explain it in the client's language?

The three metrics have names that mean nothing to most clients. Do not lead with them. Lead with what the number means in their business.

CPM rising means your ads are becoming more expensive to run. The platform is charging more to reach your kind of customer. That can happen because more competitors entered the auction, because demand in the category dropped, or because the audience you are targeting is smaller than it used to be.

CTR dropping means fewer people who see the ad are interested enough to click. The message may be wearing thin from repetition, or it may no longer be saying something they have not heard before.

CVR dropping means people are clicking but not buying or booking. The problem is not in the ad. It is in what happens after the click: the landing page, the offer, the price, the form, the checkout. The ad is doing its job. Something else is failing.

Put it in those terms. Not in metric names. The conversation sounds like this.

Instead of: "Our CVR is down 22 percent this week."

Say: "The ads are still getting clicks, but fewer of those clicks are converting on the landing page. Either something on the page is not landing the way it did before, or the offer itself needs a look. We are checking the page right now and I will come back to you by Wednesday."

The client you are talking to runs a business. They understand "fewer people are buying after they click" instantly. They do not need to know what CVR stands for. Give them the translation, not the metric.

What if it really was my fault?

Own it specifically. Vague apologies read as incompetence. Specific ownership reads as competence.

There is a real difference between "I dropped the ball on this one" and "we paused the top-performing ad set on Tuesday to test a new audience and did not reactivate it when the test underperformed. That cost us three days of volume. The original set is back on now."

The second sentence is harder to say. Say it anyway. Clients are not expecting perfect execution. They are expecting honesty and a clear path back. A specific account of what happened, followed by what you changed and what the expected recovery looks like, is the most trust-building thing you can do on a bad call.

The one thing you cannot do is let a client discover the mistake before you name it. Moxo's 2026 State of Churn report found that roughly 43 percent of B2B client churn happens in the first 90 days of a relationship. Early churn is almost always a trust problem, not a performance problem. Clients are deciding whether you are someone they can believe. Walking into a call having already named what went wrong, before they ask, answers that question clearly.

Specific ownership also limits the damage. "We made a targeting error on Tuesday" is a contained incident. "The week was just soft" with no further explanation is an open question that the client will fill in with their own worst-case interpretation.

Name the failure. Name the fix. Name when you expect to see recovery and what the recovery signal will look like. That is a complete answer.

How do I stop one off week from becoming a churn spiral?

Two things drive the churn spiral: vague communication and answering only when clients ask. Fix both.

On communication: after every significant call about performance, send a written note. Not a report. A note. Two short paragraphs. Paragraph one: what moved, in plain language, with the specific metric. Paragraph two: what you are doing about it, tied to the layer that moved, with a date you will report back.

The note does three things. It confirms that you and the client have the same understanding of what was said on the call. It creates a record the client can reread on Friday when anxiety spikes again. And it replaces the silence that usually fills the space between calls, the silence that the client's imagination populates with "they don't know what they're doing."

Swydo's analysis of why clients leave agencies found that the top two cited reasons are lack of proactive strategic guidance, cited by 68 percent of former clients, and poor communication, cited by 57 percent. Price ranked sixth, at 37 percent. An off week is not primarily a performance problem. It is a communication opportunity that most agencies miss.

During a soft stretch, increase how often you send short updates. Not full reports. Short messages. "Still watching the CTR. It ticked up slightly Tuesday. Thursday I'll send the full picture." That message takes twenty seconds to write. It prevents the client from spending three days wondering whether you noticed the dip.

Focus Digital's 2026 marketing agency churn research found that paid-advertising agencies see roughly 49 percent annual client churn with typical contract lengths of three to six months, the highest churn rate of the service specializations it measured. Retainer-based agencies that build in structured communication average around 18 percent annual churn. The difference is not always in the results being delivered. It is in whether the client feels informed between the good weeks.

The ANA/4As client-agency tenure study from 2025 measured average relationship length across agency types. Media agency relationships averaged about 3.7 years, the shortest tenure among all agency types measured. That tenure is short enough that losing a client over one bad communication cycle, rather than sustained bad performance, is genuinely the more common failure mode. The written note and the mid-week check-in are not overhead. They are what keeps clients from leaving.

Side-by-side comparison of a guessing answer ('The week was soft, we'll test more creative') versus a named answer ('Click-through rate dropped 30 percent but CPM held flat, two new angles go live Thursday').

The diagnosis method that makes this answer repeatable is in the Hub.

Free tier shows how to diagnose what actually caused the off week before the next client call.

Join the Hub free →

Frequently asked questions

Should I tell the client before they notice?

Yes. Always. Not to get ahead of the blame, but because the moment you wait for them to notice, you have let the client define what happened before you did. They will decide what the dip means before you have a chance to. Get in front of it with your assessment: which layer moved, whether it is inside normal variance, and what you are watching. A short proactive message that says "I saw CTR soften mid-week, still within range, watching it through the weekend" costs almost nothing. Waiting for the Friday Slack and then answering it under pressure costs significantly more.

How often should I report during a bad stretch?

More often than normal, but shorter. During a soft stretch, daily or every-other-day short messages on the specific metric you are watching are better than a single comprehensive report at the end of the week. The report at the end of the week is what you deliver once the picture is clear. The daily messages prevent the anxiety from compounding in the silence between reports. Keep them to two or three sentences. "CTR is holding, CVR still soft, checking the landing page today." That is enough to tell the client you are present and reading the account.

When is an off week actually a fire?

When one or more of these is true. First, the drop is large and persisting into a second week with no sign of recovery. A 10 to 15 percent dip that recovers by Wednesday is variance. A 40 percent drop that holds across two full weeks is a real break that needs a diagnosis call. Second, you identified the layer that moved but you do not have a named fix that maps to that layer. If CVR dropped and you have already ruled out landing page and offer issues and the drop is still happening, you have a problem you have not diagnosed yet. Name that honestly. Third, the client's business is directly harmed rather than the account metrics. If the account's performance decline is pushing them below break-even on their ad spend, that is a different conversation than a soft week inside a profitable account. The urgency and the framing both change. Treat those three conditions as the threshold. Below it: the communication approach above. Name the layer, send the written note. Above it: escalation call, clear diagnosis, and a recovery plan with timelines.

Related reading: Why did my ads stop working? · 7 things to check before testing more creative · What the Realignment Hub teaches