The message usually arrives late in the evening, through the app's chat or a direct message on social media: "Found this in my food. Refund me and I won't post it." Attached is a photo — a hair, a piece of plastic, something that looks like mould — and, implied or explicit, a one-star review with the same photo waiting to be published.

This is extortion, and it works because the maths is so lopsided. A refund costs you the price of a meal. A one-star review with a photo of "your" food costs you customers you will never know you lost. Most owners pay, and the person on the other end knows it.

Here is how to handle it so that you neither pay for a lie nor hand a liar a stage.

First: is the photo real?

Everything depends on this, and it is the step most owners skip because the threat feels urgent. It is not. A review takes seconds to post and days to remove; an extra hour to check the photo costs you nothing.

Get the original file if you can (download from the chat rather than screenshot), and run through the checks in our guide to spotting edited photos: lighting that does not agree, a defect that is sharper or blurrier than the food around it, packaging that is not yours, a reverse-image search. If you have FraudBite, upload it and read the score, the heatmap and the metadata notes.

If the photo is genuine — and sometimes it is — you have a normal complaint with an unusually aggressive customer. Apologise, refund, fix the process, and let the review land if it must; a fair response to a real problem does you no harm.

If the photo is not genuine, you have leverage, and you should use it calmly.

Do not pay under threat

Paying a refund to prevent a review establishes that it works, and the same person will be back — or their friends will. It also does not guarantee anything; the review often gets posted anyway.

Reply once, in writing, in the same channel, and keep it to facts:

Thanks for getting in touch. We take every complaint seriously, so we've looked at the photo you sent carefully. It doesn't appear to show the order as it left our kitchen (we photograph every order at the pass), and the image shows signs of having been edited. We're happy to look again if you can send the original photo from your phone.

No accusation, no threat, no mention of reviews. You are not calling them a fraudster; you are declining to refund and asking for better evidence. Save the whole conversation.

If the review goes up

Report it to the platform, with your evidence. Google, Tripadvisor and the delivery apps all have a process for reviews that violate their policies, and extortion and fake content are explicit violations on all of them. Your report is far more likely to succeed if it contains the chat where the refund was demanded in exchange for not posting, plus your image analysis. Removal is not fast and not guaranteed, but it happens more often than owners expect when the report is specific.

Reply publicly, briefly, to the audience — not the reviewer. Everyone reading a one-star review is looking at your response to decide whether to believe it. A good response is short, calm, and factual:

We're sorry you feel this way. We checked this order carefully: our pass photo shows it as it left the kitchen, and the image attached to this review shows signs of editing. We asked for the original photo and didn't receive it. We'd still welcome the chance to put things right — please contact us directly.

That paragraph tells future customers three things: you take complaints seriously, you check, and you are not a pushover. It does not name-call, because the audience punishes that.

Do not post the heatmap in the reply. Keep the analysis for the platform's report and, if it comes to it, for your lawyer. Public forensic arguments read as defensive.

The repeat pattern

Review blackmail is rarely a one-off. Watch for the same photo appearing under different names, the same phrasing ("found this in my food"), and threats that arrive right at order value thresholds. A note in your order system against the address or phone number is enough; a pattern documented over three incidents is something a platform's account manager will act on where a single complaint gets ignored.

When to involve someone else

If the demands escalate — repeated threats, demands for money beyond the order value, threats to contact health authorities with fabricated evidence — this stops being a customer service issue. Keep everything in writing, do not engage further in the chat, and take the file to your local police or a solicitor. Extortion is a crime in every jurisdiction we operate in, and a documented image analysis plus the threat itself is exactly the file they need.

The habit that prevents most of it

Photograph the order at the pass. It takes two seconds, and it means that every threat you receive is already answered before it arrives. Combined with a quick image check, it turns "refund me or else" from a dilemma into a form letter.

Want to see what the check would say about a doctored complaint photo? Try it on one of our samples — thirty seconds, no account.