Ending a hard call with next steps

The last ninety seconds decide whether the client hangs up with an action or with a worry.

The idea

In a difficult call, the middle is explanation. The end is the only part the client can use. People forget your reasoning within a day; they remember what they were told to do, and when they will hear from you next.

So the close is not a wind-down. It is four deliberate moves: say the decision back, name who does what by when, say out loud what you still don't know, and put the next contact in the diary. Together they take about ninety seconds.

Cut any one of them and the client fills the gap themselves — usually with the worst reading available.

the last ninety seconds

Halden Foods · analytics error call

You are the account lead. You have just told Priya, the marketing director, that three weeks of her conversion data were mis-tagged. Scrub through the closing minute. Toggle the four moves and watch how she leaves the call.

Press play, or step through. The hard news is already delivered — what follows is everything Priya can act on.

position0:00

moves in4 / 4

close length90s

time saved0s

clear

to: priya.raman@haldenfoods.example · sent 14:32, four minutes after the call

Priya,

— Sam

How it works

Four moves, in this order, in roughly ninety seconds. The order matters: the decision anchors the actions, the actions expose the gaps, the gaps make the next contact necessary.

Restate the decision in one sentence. Not the reasoning — the conclusion, in words the client could repeat to their own boss. Then check it: “Does that match what you heard?” If they correct you, you have just avoided a week of divergence.
Name who does what by when. Every action needs a person, a verb, and a date — including the client's own action. A close that lists only your tasks lets the client believe they are waiting on you while you are quietly waiting on them.
Say what is still unknown. Out loud, before they ask, with a date for the answer. An unknown you name is information; the same unknown discovered later is a second failure on top of the first.
Book the next contact. A specific time, and the invite sent before you hang up. “I'll be in touch” puts the next contact inside your memory. A calendar entry puts it outside it, where the client can see it.

The whole thing fits in the time people spend saying goodbye badly:

closing budget — Halden call
------------------------------------------
1  restate the decision            17s
2  who does what by when           25s
3  what is still unknown           15s
4  book the next contact           15s
                                   ----
   the four moves                  72s
   framing (apology, thanks, air)  18s
                                   ----
   full close                      90s

drop move 3 to "keep it positive":
   close = 18 + 17 + 25 + 15     =  75s
   saved                          =  15s
   cost = Priya finds the February
          error on her own dashboard,
          and now doubts the March fix too

When to use it

Situation	How it lands
Service recovery after an error your side caused	Use all four. This is the case the structure was built for — the client's trust is rebuilt by the close, not the apology.
Bad news where no decision has been made yet	Use a reduced close: name the unknown, name who is deciding and by when, and book the contact. Do not invent a decision to sound in control.
Escalation while the client is still angry	Wait. The four moves only land after the person feels heard. Closing early reads as “I'd like to end this call”.
Routine check-in or discovery call	A light version. The “decision” may just be we'll scope it — that is still worth saying back.
Email or chat only	Same four moves, but they belong in the first five lines, not the last. Nobody scrolls to find their own deadline.

The trade-off: a structured close costs ninety seconds and can feel abrupt if you deliver it as a checklist. Say it in your own voice, in whole sentences, and pause after the decision so the client can push back.

Watch out for

Restating your preferred version of the decision. The close is a checkpoint, not a second chance to shade the outcome your way. If you soften “we're pausing spend” into “we're reviewing spend”, the client agrees to a different thing from the one you'll do on Monday.
Owner-less verbs. “We'll get that fixed”, “someone will look at it”, “it's with the team.” None of these is an owner. Name a person, a date and — if you're cross-region — a time zone. And say the client's action last, so it's the one they hang up holding.
Hiding the unknown to sound competent. The instinct to end on good news is strong and it is expensive. Unnamed unknowns come back as surprises, and a surprise costs you trust twice: once for the problem, once for the withholding.
A next contact with no time on it. “I'll come back to you next week” is not a booking. It also quietly transfers the chasing to the client, which is exactly the work you are trying to take off them.
New information in the final thirty seconds. If a thing is not a decision, an owner, an unknown, or a date, it belongs in the next conversation. Ending on a fresh complication undoes everything the close just built.

Worked example

Asked in interview: “Tell me about a time you had to deliver bad news to a client.” The strong answer spends most of its time on the last ninety seconds, because that is where the judgement lives.

“We'd mis-tagged Halden's checkout flow, so three weeks of conversion data were inflated. I called their marketing director, walked her through what happened and apologised in the middle of the call, not the end. Then I closed it deliberately. First the decision: we pause paid spend on the two affected campaigns until Thursday, and nothing restarts before the re-tag. I asked her whether that matched what she'd heard, and she added one campaign I'd have missed.

Then owners: I'd send corrected March figures by Wednesday 5pm, our engineer would re-tag on Tuesday morning, and her developer needed to approve the container push by Tuesday noon — that last one mattered most, because without saying it she'd have assumed the delay was entirely ours. Then the thing I didn't know: whether February was affected too. I said I'd have the audit answer Wednesday and would tell her either way. Then I booked Thursday at nine and sent the invite before I hung up.

The email went out four minutes later with the same four blocks. She replied with one line: ‘Clear — dev is briefed.’ We held Thursday. February turned out to be clean, and because I'd flagged it, telling her that was a good-news call rather than a confession.”

Check yourself

1. The client asks when they'll know about February, and you genuinely don't know. Which closing line is strongest?

2. Which of these is a real ownership line?