On 16 August 2026, a reader named Fechitor emailed support [at] bigmacindex.app to tell us that a sentence on this site was false. He has family in Tabriz. They are certain there has never been a McDonald’s in Tabriz — or anywhere else in Iran — and he wanted to know what “sources” we thought we had.

He also told us how he’d found us. He had asked Google whether McDonald’s ever operated in Iran. Google’s AI answered yes, and cited this website as the evidence.

He was right. It was worse than he knew. We apologise to him and to every reader who took that sentence at face value, and we thank him for writing in. The claim came down from all six places it appeared, on the same day, and the corrected Iran page now carries a dated note explaining what happened.

This is the full account: what we published, exactly when, what we found when we finally checked it, and where the claim actually came from. That last part is not what either of us expected.

The timeline

18 May 2026 — we publish the Iran page. Its opening paragraph asserts a pre-1979 McDonald’s in Tabriz, sourced to two secondary articles. The same claim goes into the page summary, the meta description, and a FAQPage block of structured data.

Three months pass. Google begins answering “did McDonald’s operate in Iran” with yes, citing this page.

16 August 2026, morning — Fechitor’s email arrives.

16 August 2026, same day — we re-read our own citations, find the problem described below, and remove the claim from all six locations: the blog post, the country page body, the FAQ structured data, two cross-references on other country pages, and a source-code comment instructing future editors to keep it.

The sentence that ran for those three months was:

McDonald’s operated very briefly in Iran during the 1970s — most sources point to a Tabriz location around 1971 — before the 1979 Iranian Revolution ended Western brand presence in the country.

Two secondary articles were cited. Neither, it turns out, was load-bearing.

What we found when we checked

One of the two sources we cited does not contain the claim.

We linked a Food Republic piece on countries where McDonald’s is banned. Read in full on 16 August 2026, it does not mention Tabriz. It does not mention 1971. It does not say McDonald’s traded in Iran before 1979. It says the chain has been unwelcome since the revolution and that a 1994 franchise attempt was destroyed within days — both of which we could have written without the invented part.

That’s not weak sourcing. That’s attaching a claim to a publication that never made it — and then linking to it, so readers could see for themselves how well sourced we were. The second source, Atlas Obscura, now blocks automated reading, so we could not re-check it.

We never had a primary source, and we said so in the wrong place. The old page contained this line, three paragraphs down: “The exact number of outlets is unclear and I haven’t been able to nail down a primary source.”

It also contained a [TODO: verify] note — visible in the published HTML, to anyone, for three months. That’s the kind of thing you find when you go looking and stop being pleased with yourself.

Hedges don’t travel

Here’s the part I’d want another publisher to take away.

The caveat sat in paragraph three. The assertion sat in the first sentence, the summary box, the meta description, and — this is the one that mattered — a FAQPage block of structured data that this site publishes for search engines to read directly.

Structured data is not prose — it’s a machine-readable answer to a machine-readable question, stripped of every sentence around it. Ours said:

Q: Did Iran ever have a McDonald’s? A: Briefly. Most sources point to a Tabriz location around 1971.

No hedge. No “unclear.” No “I couldn’t find a primary source.” Those words existed — several hundred words away, in a format nothing downstream reads.

Fechitor found us because Google’s AI was citing this page as evidence and answering yes. It wasn’t misreading us. It was reading the part of us that was built to be read.

If your caveat isn’t in the first sentence and in the schema, you haven’t published a caveat. You’ve published a claim with a private note attached.

Where the claim actually comes from

Fechitor’s theory was that it started as a Wikipedia hoax, repeatedly inserted and repeatedly removed. That’s a reasonable guess, and it’s the one most people would make. It turns out to be backwards, and the real answer is more uncomfortable.

The English Wikipedia article List of countries with McDonald’s restaurants carries this note:

“Sources such as NY Times, LA Times and the Telegraph have claimed that McDonald’s operated stores in Iran prior to 1979, but no source has provided evidence for this claim.”

“The number of McDonald’s outlets in pre-1979 Iran, or even the existence of such outlets, remains unclear.”

Iran appears in neither of that article’s tables — not under current locations, not under former ones.

So the chain runs the other way — and it starts a lot further up than a wiki. Three newspapers of record printed it. Wikipedia’s editors caught it, refused to put Iran in the table, and left a standing annotation saying the sourcing does not exist. Then secondary food-and-travel sites picked it up from the newspapers. Then we picked it up from them. Then Google’s AI picked it up from us.

Wikipedia is the layer everyone in this business is trained to sneer at. On this specific claim it was more careful than three papers of record, more careful than the aggregators, and more careful than us.

Two things I have to flag about that paragraph. I verified Wikipedia’s text myself, on 16 August 2026. I did not independently verify that those three newspapers made the claim — the LA Times blocks automated fetching and the Telegraph sits behind a paywall. What I’m reporting is Wikipedia’s characterisation of them, clearly labelled as such, because doing otherwise would repeat the exact mistake this article is about.

What we are not claiming

We are not claiming McDonald’s never operated in Iran.

We can’t — and neither can anyone else writing about this from a desk. Absence of evidence for one small franchise in 1970s Iran is not proof that none existed, and the honest position is narrower than either side of the argument wants it to be: no contemporary reporting, company record, or franchise filing supporting it has surfaced, and this site should not have been repeating it as fact.

There will now be a wave of pages confidently stating that McDonald’s never operated in Iran. That’s the same disease pointing the other way. If you hold a primary source — a 1970s advertisement, a company document, a photograph with provenance, a franchise filing — we would like to see it, and we’ll publish it whichever way it cuts.

Fechitor’s family testimony is evidence, and it’s what made us look. It isn’t what settled it. What settled it was reading our own citation and finding nothing there.

What we can actually source

Here is the thing that makes the whole episode sillier. There is a well-documented McDonald’s story in Iran. It just isn’t the one we ran.

Tehran has imitator chains. The best known is Mash Donald’s, run by an owner named Hassan, who decorates the shop with reworked golden arches and explained the name to NPR in 2013:

“If I had called my restaurant McDonald’s, I’d get a visit from the hard-liners. My son advised me to go for Mash Donald’s. It sort of sounds the same.”

That is a named person, on the record, to a named outlet, with a second outlet reporting the same shop. It is more evidence than the entire Tabriz claim ever had — and it was sitting on our own Iran page the whole time, three headings below the sentence that was wrong. There are others — Pizza Hot for the obvious target, and straightforwardly Iranian chains like Tehran’s Burgerland that don’t imitate anyone.

For a site that publishes prices, the knockoffs are also the only thing resembling a data point. Iran has never appeared in The Economist’s survey, so any Iranian figure here would be a shadow reading off a substitute product — and our Iran page explains why we don’t publish one. Two reasons, both fatal. Iran runs an official rial rate alongside a widely used unofficial market rate, and locals quote prices in tomans (one toman being ten rials), so the same menu price converts to wildly different dollar figures depending on which rate you pick. And a Mash Donald’s burger is not a Big Mac. The index needs the same object everywhere; that’s the whole point of it.

So we have a sourced story about imitation chains, a currency that can’t be converted cleanly, and a product that isn’t the reference product. Three honest, interesting, publishable facts — and we skipped past all of them to assert a fourth thing that nobody could verify.

The uncomfortable part

This site exists to argue that a single published number is a claim about how the number was made, and that you shouldn’t trust one without the other. That’s what the methodology page is for. It’s why the country pages mark which tier a figure comes from.

And then we ran an unsourced sentence about Iran for three months, in the schema, in the meta description, at the top of the page.

I don’t have a clever way out of that. The lesson isn’t “be careful” — everyone is careful, and everyone believes it about themselves right up until a reader emails. The lesson is that the checks have to run on the parts of your site that machines read, because those are the parts that leave the building.

Readers are the error-correction mechanism

This is the third time a reader has fixed something here that our own process missed. A submission from Egypt corrected a price. A till receipt from Pas-de-Calais checked a French figure against an actual menu board — a photograph of a real receipt, from a real till, in a town of eleven thousand people. Now Iran.

Not Google. Not the newspapers. Not our own process. Three people who happened to know something true, and spent twenty minutes of their own evening writing to a stranger about it.

I want to be plain about this rather than clever. I run this site. I check things, and I still shipped an unsourced sentence about a country I have never been to, and left it up for three months, in the one format built for machines to quote. The reason it is fixed today is not that our process worked. It’s that somebody cared enough to write in — and got nothing out of it except a better page on somebody else’s website.

Reader reports have been a formal data tier here for a while, with their own file and their own rules. After today they are also the only part of this pipeline that has never once been wrong.

So: thank you, Fechitor. And thank you to the reader in Egypt, and to Finn in Pas-de-Calais who photographed a receipt for us. Corrections from readers are the only thing that has reliably made this site more accurate, and every one of them has made it better than I could have on my own. Please keep sending them — including the ones that sting.

Corrections to this or any page: contact us. We publish them with the date and, unless you’d rather stay anonymous, your name.