When official tourism figures change: provisional, revised and final numbers

Provisional means the publisher may still change the number; revised means it already has; final means the series is closed. In Türkiye the Ministry's border bulletins are marked provisional, and TÜİK updates a reference year quarter by quarter. Quote publisher, date and status every time.

You quote a visitor figure in an October board paper. In January the same table shows a slightly different number, and nobody has done anything wrong. The figure moved because the publisher said it would. Understanding revised tourism statistics in Türkiye means reading the labels publishers put on their own numbers, not the news stories written about them. After this article you will know which figures can move, why, and how to quote them so a revision never embarrasses you.

Why do published tourism numbers change after publication?#

Because a published figure is a snapshot of a process that keeps running: late returns arrive, survey samples are refreshed, and classifications are aligned after the first release.

A facility files late. A border gate sends a corrected file. A survey response arrives after the deadline. The first release was correct for what it had. When the missing pieces land, the publisher incorporates them and the number moves.

Sometimes the estimate itself is rebuilt. When a survey component is re-estimated with better information, the estimate built on it moves, and so does every ratio calculated from it. Occasionally a coding or classification mistake is fixed; this is the only case where the earlier number was actually wrong. And sometimes a category is reclassified, so the series moves without anything changing on the ground. That kind of change is announced in metadata, rarely in the headline.

Revisions are a sign of a working statistical system, not a failure. The problem starts only when the label is dropped and a provisional number is quoted as settled. Each publisher sets its own revision policy, and that policy is the contract with the reader: what will be revisited, when, and under which label.

Which official Turkish tourism figures are marked provisional?#

The Ministry of Culture and Tourism's border bulletins say so on their own sheets.

The September 2026 border bulletin marks its tables with the note "Veriler geçicidir", meaning the data are provisional. That note, not the news story written about it, is what you cite. In that bulletin, 40,938,498 foreign visitors arrived in January to September 2026, 1.52% fewer than a year earlier, and total visitors including 8,193,564 citizens resident abroad were 49,132,062, 1.72% fewer (Ministry of Culture and Tourism, September 2026 border bulletin).

One mechanism is worth knowing. The bulletin's total-visitor table is updated after each quarterly TÜİK bulletin, so the total column can move for a reason that has nothing to do with the border count itself. Ministry facility stock figures are published as provisional. Cite the border bulletin as "September 2026 border bulletin".

None of this makes 2026 a growth story. The year is flat, and a provisional count does not change that. For how the underlying counts are built, see how Türkiye's border statistics are built and how the monthly accommodation statistics are made.

Public data. No endorsement implied.

How does TÜİK revise a tourism year?#

Quarter by quarter, not in one annual correction: TÜİK revisits a reference year as later information arrives, most visibly as health spending is refreshed from its Health Tourism Survey.

The worked case: second-quarter 2026 visitor revenue was first published on 31 July 2026 as USD 15,655,868 thousand and appears marked revised at USD 15,661,610 thousand in the table released with the third-quarter bulletin (TÜİK, 30.10.2026). That is the normal shape of a revision: real, documented, and irrelevant to any decision a hotel makes.

Read the marker, not the memory. An (r) in a TÜİK table means revised, and the metadata note tells you which revision round you are looking at. TÜİK publishes quarters, so four quarters added together is your arithmetic, not an official figure. A chart that adds quarters and labels the result "2026 so far" presents a derived number as if the publisher had issued it. For the vocabulary around such comparisons, see base effect, real, nominal and points explained.

What changes most, and what rarely does?#

Counts move least, estimates move more, and ratios of two estimates move most; facility stock moves slowly and by administrative decision.

Border arrivals come from administrative border records. Departing-visitor counts come from TÜİK's survey of departing visitors. The two move for different reasons and at different sizes. Revenue and spend also come from a sample survey. When the sample or one component such as health spending is refreshed, the estimate built on it moves.

Derived figures such as spend per visitor or nights per trip combine two estimates, so the numerator and the denominator can both move at once. Facility stock and certified capacity change when a certificate is issued or cancelled; that movement is administrative, not statistical, and it is dated. A reclassification can shift a series further than any data refresh, and it usually appears in the metadata rather than in the headline number.

The practical consequence: the more derived the figure, the more of it you should write as "about", and the more carefully you should keep the version you used.

What should you quote: the exact figure or "about"?#

Quote the exact figure with publisher, date and status when the number is the point; use a rounded figure with "about" when it is background, and keep the source line either way.

The three-part citation is publisher, publication date, status: "Ministry of Culture and Tourism, September 2026 border bulletin, provisional." Precision is a signal of checkability. A rounded number with no source line cannot be checked by a reader and cannot be corrected by you later.

Rounding a provisional number into a confident claim is the commonest error. "About 41 million" is honest; a bare "41 million" lifted from a provisional table is not. Keep the version you used: file the bulletin or note its release, so that a later revision can be traced back to the sentence it changed.

One number, one source. Never place TÜİK's departing-visitor count and the Ministry's border arrivals in the same sentence as though they measured the same moment. If two sources disagree, say so and say why they measure different things. Do not average them into a third number that no publisher stands behind. For the broader discipline, see how to read and cite official Türkiye tourism data.

What do you do when a number you have already used is revised?#

Say so in one line and move on: a revision note protects the reader's trust and costs a sentence.

Do not silently swap the figure. A reader comparing two of your reports will find the difference and will not know which version to trust. A revision note has three parts: what changed, by how much, and whether the conclusion changed. Most revisions do not change a decision, and saying that explicitly is more useful than the new number on its own. If the revision does change a decision, say that too, and say what you will do differently as a result.

Keep a version log for standing reports such as monthly revenue packs and board decks, so the same figure is never quoted with two different dates. Check the calendar before you publish: knowing which release is provisional and when the next one lands prevents most of these notes from being needed. The calendar of Türkiye tourism data releases lists what comes out when.

What to do next#

Next time a tourism figure is about to enter a document, open the source file first and read its label, then write publisher, date and status beside the number. The three ideas fit together: a published figure is a snapshot, the label is the publisher's contract with you, and a citation that carries all three parts survives any revision without embarrassing anyone.