Writing to a client without creating a problem
Clients rarely leave because something went wrong. They leave because they found out late, from the wrong side.
Most client damage is not caused by the work. It is caused by three or four sentences written quickly, at the wrong moment, in the wrong tone.
Promise the date you would bet on
The instinct under pressure is to name the date you would hit if everything went perfectly. That date is a wish, and the client hears a commitment.
Give the date you would bet money on, which is usually the optimistic one plus half again. Delivering early is a small pleasant surprise and costs nothing; delivering late is a conversation that spends credibility you will need later for something that actually matters.
Bad news travels best early and plainly
Nobody wants to send the message saying it will be late, so it gets postponed until there is more information — and by then the client has planned around a date that was never real.
Send it as soon as you believe it, not when you can prove it. Three parts, in this order: what changed, what it means for them, what you are doing about it. No preamble, no explanation of the internal reason before the fact itself. People forgive delays they were told about in time and remember delays they found out about themselves.
Silence is read as bad news anyway
When a client hears nothing, they do not assume everything is fine. They assume the worst version and rehearse it, so by the time you write, you are answering an argument you did not know was happening.
Even with nothing to report, send two lines on the day you said you would. "Still on track, next update Thursday" costs thirty seconds and prevents the entire spiral. Predictability is most of what people mean when they say a supplier is reliable.
Write it, then remove the defensiveness
Read the message back once before sending and take out the parts explaining why it was not your fault. They rarely persuade and they always signal that you expected to be blamed.
State the situation, the effect and the next step. If something genuinely was outside your control, one clause is enough — a paragraph of it turns a factual update into an argument the client had not started.
Questions people actually ask
What date should I give a client?
The one you would bet money on, usually the optimistic estimate plus half again. Delivering early is a small pleasant surprise; delivering late spends credibility you will need later for something that matters more.
When should I tell a client about a delay?
As soon as you believe it, not when you can prove it. Three parts in order: what changed, what it means for them, what you are doing. People forgive delays they were told about and remember the ones they discovered.
What if I have nothing to report?
Send two lines anyway on the day you promised. Silence is read as bad news — clients rehearse the worst version — and “still on track, next update Thursday” costs thirty seconds.
