Direct answer

A refund demanded with a review threat is two problems. Handle them separately.

A refund demanded under a review threat is two separate problems: whether any money is honestly owed back, which is a quality question, and the threat itself, which is a pressure tactic. Handle them separately, in writing, and never trade a refund for a review change. If a step is genuinely wrong, fix it because it is owed. If the work matches what was agreed and the client accepted it, the threat changes nothing about the money. Here is the full sequence, in order.

See the instrument first: the live demo runs one real call and hands you its verifiable receipt, no signup.

Updated October 21, 2026. The two delivery records quoted below are real and public, with their states read back from the live service.

Step 1. Slow down and move everything to writing

Do not answer a threat in the heat of the message that carried it, and do not answer it on a call. Move everything to writing and keep it there. Save the threat exactly as it was sent: the full message, its date, and its exact wording, in a screenshot and in plain text. That saved wording matters more than your summary of it, because platforms and review sites treat a refund-for-silence demand as extortion-like behavior, and a report stands or falls on the wording itself. A message that says refund me or I post a one-star review is a different document from a message that says I am unhappy with the work. Save which one you actually received before you reply to either.

Step 2. Judge the work on its merits, against the record

Set the threat aside and judge the work cold, against two documents: the agreement and the acceptance record. What was promised, what was delivered, and did the client check and accept it? If a step is genuinely wrong, fix it or refund that part, because it is owed, not because of the threat. If the work matches the agreement and the record shows acceptance, there is no honest refund to give, and paying one anyway does not settle a quality question. It settles a pressure question, on the wrong side. The two real records below show how different those two positions look in practice:

Accepted: the client checked the work and released it

Pending: the honest middle state, no decision yet

Notice what neither card claims. Neither says the work was good. They say what was delivered, how many steps were checked, and what the client decided, or has not decided yet. That is exactly the ground a refund-under-threat conversation should stay on.

Step 3. Reply once, short and boring

Send one reply that does three things and nothing else. Name the record: what was delivered, when, and what the client accepted. State what you will fix if anything is actually wrong, and ask the client to name the specific part that fails the agreement. And do not mention the review at all. Naming the threat in your reply turns the exchange into an argument about the threat, invites a denial, and gives the client a fresh thread to perform in. A boring factual reply gives them nothing to push against except the record itself. Then stop. One reply said once is the position. Ten replies defending it is a negotiation.

Step 4. Never trade money for a review

Never offer money, discounts, or free work in exchange for removing or softening a review. Review platforms forbid that trade, so the offer itself can become a second policy problem on your side of the record. And even where no platform is watching, the trade converts one threat into a standing arrangement: a client who is paid once for silence has learned the price of silence, and the next disagreement arrives with the same demand attached, usually larger. If money moves, it moves because a step failed the agreement, in writing, for that stated reason only.

Step 5. If the review lands, report it, then answer it once

If the review is posted, do two things in this order. First, report it to the platform with the saved threat attached. The saved message, with its date and exact wording, is the report. A report that says the client threatened me, without the message, is an opinion. A report carrying the message itself is evidence, and platform teams act on evidence. Second, post one factual public reply built from the delivery record. Write it for the future readers who are the real audience: they will never meet the reviewer, but they will read your reply and decide what kind of seller you are. Short, calm, checkable facts only: what was delivered, when, and the acceptance state. Then leave it alone. Do not edit it daily, do not reply to every follow-up comment, and do not let the public thread become the negotiation the client wanted.

Step 6. Prevention: deliver as a record, so the next threat meets acceptance

The durable fix is to deliver as a record with acceptance, so the next threat meets a client who already accepted the work in writing. When each step is saved as it happens and the client checks and accepts it on one link, a later claim that the work was worthless has to argue with that recorded acceptance, its date, and its release state. Sellers without that record end up arguing from memory and message threads, which is the argument a threat is designed to win. Put the acceptance step in the written agreement, keep final payment after acceptance where the job allows it, and the threat loses most of its leverage before it is ever sent.

What the record does not do

A verifiable receipt proves the saved result was not changed since it was saved. It does not prove the work was good, and it cannot stop anyone posting a review. Platforms decide what stays up, and no record on your side changes that. What the record does is narrower and still valuable: it keeps your reply factual instead of emotional, and it gives a platform report something solid to stand on next to the saved threat. This page is general information, not legal advice. If a threat involves large sums, safety, or a formal legal claim, get advice for your situation and your place.

Frequently asked questions

My client is threatening a bad review unless I give them a refund. What should I do?

Keep the two problems separate. One is whether any money is honestly owed back, which is decided by the agreement and the delivery record. The other is the threat itself, which is a pressure tactic and gets no reward. Save the threat exactly as it was sent, with its wording and date. Check the work against the agreement and the acceptance record: if a specific step is genuinely wrong, fix it or refund that part because it is owed. Then reply once, short and factual, naming what was delivered, when it was delivered, and what the client accepted, and offer to fix anything that specifically fails the agreement. Do not mention the review in that reply, and do not offer money, a discount, or free work in exchange for removing or softening a review. If a review appears, report it to the platform with the saved threat attached, then post one factual public reply built from the delivery record for future readers. This page is general information, not legal advice.

Should I ever refund a client who is threatening me?

Yes, but only when the refund is owed on the merits, never because of the threat. If a step fails the agreement or was never delivered, fix it or refund that part promptly and say plainly that you are doing so because it is owed. If the work matches the agreement and the client accepted it, a refund paid to stop a review teaches that threats work and invites the next one. Decide the money question as if no review existed, decide the threat question separately, and keep both decisions in writing.

Is threatening a bad review unless I refund against the rules?

On most major review and marketplace platforms, yes. Asking for money or a refund in exchange for removing, changing, or not posting a review is generally treated as extortion-like behavior or review manipulation, and platforms usually provide a report path for it. Rules differ by platform, so read the policy for the site involved and keep the exact message, because a report stands on the saved wording, not on your summary of it. A platform decides what stays up, and a clear saved threat plus a delivery record gives that decision something solid to stand on.

What should I write in a public reply if the bad review is posted?

Write for the future reader, not for the reviewer. Keep it short, calm, and built only from facts a stranger can check: what was delivered, when it was delivered, and whether the client accepted it. Name one offer to fix any specific part that fails the agreement, and stop there. Do not argue, do not accuse the reviewer of lying, do not mention the threat unless the platform report is already filed and the wording stays factual, and do not post private details, messages, or payment information. One boring factual reply does more for the next client reading it than any argument you could win.

How do I stop a threat like this working on the next job?

Deliver as a record with acceptance, not as loose files and messages. When the work runs through Zambo, every step is saved as it happens and the client checks and accepts or flags it on one handover link, so a later claim that the work was worthless has to argue with the client acceptance already on file. For $0.99, deliver with an acceptance record so a threat meets an accepted record, not a memory: get the Day Pass. A Day Pass covers 24 hours, no subscription. Ongoing use is the Zambo Pass at $29 a month or $290 a year, and $19 a month for $ZAMBO holders, which adds private handover links, your branding, and the approval record as PDF and JSON files.

Related