Türkiye tourism by year: reading the long series without its traps

Türkiye's long tourism series come from two publishers: TÜİK's table of visitor income and average nights, 2012 to 2026, and the Ministry of Culture and Tourism's yearly border bulletins from 2010. Read them separately: a 2024 revision reset TÜİK's series at 2012, revenue is nominal dollars, and 2020 is not a normal year.

A year-by-year table of Türkiye tourist numbers by year looks like a row of plain facts. Put three such tables side by side and the same year shows three different figures, because the series behind them changed while the headline stayed the same. After reading this you will know where the official long series live, which table answers which question, and how to quote a multi-year change without borrowing a trap.

Where are Türkiye's tourist numbers by year actually published?#

Two official publishers keep a long series, and they count different things: TÜİK counts departing visitors and what they spent, the Ministry of Culture and Tourism counts border arrivals.

TÜİK publishes tourism statistics (opens in a new tab) quarterly, about 30 days after each quarter, with no seasonal adjustment. Its third-quarter 2026 bulletin, No. 58144, released on 30 October 2026, carries the table to use for revenue and length of stay: "Tourism Income from Visitors, Tourism Expenditure and Average Number of Nights, 2012-2026". It gives income, spending and average nights per departing visitor, year by year.

The Ministry's yearly border bulletin is compiled from passport police records and TÜİK data and appears at the end of the first half of the following year. The 2025 bulletin carries foreign arrivals from 2010 to 2025, with citizens resident abroad listed separately. If you want the mechanics behind these counts, see how Türkiye counts its tourists at the border.

The units are not interchangeable. TÜİK's 2025 count is 63,917,057 departing visitors including 11,175,394 citizens resident abroad; the Ministry's 2025 count is 63,941,261 border arrivals, of which 52,775,055 were foreign. Neither publisher explains that small gap beyond departures versus arrivals, so do not reconcile the two totals yourself and never put one series' figure next to the other's in the same sentence.

Both files are free to download, and both carry a release date. A table copied without its release date is not a source, it is a screenshot.

TÜİK's table lives at veriportali.tuik.gov.tr, released with bulletin No. 58144 on 30 October 2026. The Ministry's bulletin lives at yigm.ktb.gov.tr, published in the first half of the following year.

Why do older printed figures no longer match the TÜİK table?#

Because after a 2024 revision TÜİK recalculated the tourism series back to 2012, so a figure printed before that revision belongs to a different calculation, not to a different year.

The revision changed the whole back series, not just the newest year. Any 2019 or 2015 number you find in an old report may simply not exist in the current table. The practical test is the endpoint: the 2025 values published in TÜİK's annual bulletin of 30 January 2026 and the same values in the 30 October 2026 table match, with no revision mark. That is what a stable endpoint looks like.

The rule that follows: quote only the current table, and name its release date in the sentence. If a reader cannot open the same table and find the same number, the citation has failed. For a fuller method, see how to read and cite official Türkiye tourism data.

Do not splice. Joining a pre-2012 figure from an old printout to a post-revision figure produces a growth rate that measures the revision, not tourism. If you genuinely need a year before 2012, ask TÜİK whether a comparable series exists rather than assuming one does. If you meet a revision mentioned anywhere, confirm it with TÜİK before you repeat it.

What does a tourism revenue series in dollars actually measure?#

It measures nominal US dollars received in a year, not value: the series is not adjusted for US inflation or for exchange-rate movement, so a higher number can mean more visitors, more spending per visitor, or simply a weaker lira against the dollar.

Definition first. TÜİK's 2025 total tourism revenue was USD 65.23 billion, of which USD 64.45 billion came from visitors and USD 782 million from transit passengers. The long table's income column is the visitor part, so quote the part you mean. For what sits inside that total, see what TÜİK's 2025 tourism revenue figure includes.

The table gives three columns that do the interpretive work: income, number of departing visitors, and average nights. Read them together or not at all. In 2012 the table shows USD 35.72 billion of visitor income from 36.46 million departing visitors, USD 980 per visitor, 12.53 nights on average. In 2025 it shows USD 64.45 billion from 63.92 million, USD 1,008 per visitor, 10.66 nights.

Source: TÜİK, table "Tourism Income from Visitors, Tourism Expenditure and Average Number of Nights, 2012-2026", released 30 October 2026. The 2019 per-visitor and nights values are left out because this article quotes only the values given in its fact list.

Those two endpoints show the shape a long series often has: many more visitors, moderately more dollars per visitor, fewer nights each. That is a mix change, and the table records it without explaining it. The table does not say why anything moved. Reasons need other evidence: source-market shifts, prices, air capacity, events. Do not read a cause into a column.

Write the nights figure exactly as published, 10.66 nights, and never round it to 10.7 or turn nights into days.

Where does the long line break, and how should you show it?#

There are two kinds of break and they need different treatment: a real shock, where the measurement is intact and the year was simply bad, and a measurement break, where the numbers themselves do not cover the same period.

A shock year: Ministry foreign arrivals were 36,244,632 in 2015 and 25,352,213 in 2016. Nothing in the bulletin's method notes marks a change in counting for those years, so a comparison across them is legitimate as long as you say what happened that year.

A measurement break: TÜİK's table gives 2020 as USD 15.17 billion of visitor income and 15.83 million departing visitors, but the two figures do not cover the same period. The mechanism: because of the pandemic the border survey could not be run in the second quarter of 2020. That quarter's income, USD 651 million, rests only on interbank card data, and the table gives no visitor count for it. So the 15.83 million is the sum of the first, third and fourth quarters. Dividing one by the other produces a ratio of two different periods, which is why any 2020 income-per-visitor figure should be left alone.

How to show a break: mark it on the chart with a note, do not draw a trend line through it, and do not compute a growth rate that spans it as if the middle year were normal. The same caution applies to any year where a publisher changes method, coverage or timing. Check the metadata before you treat a dip as demand.

How do you quote a change over years honestly?#

Use the same series, the same definition and two dated endpoints, and say out loud that the dollars are nominal.

Same series: both endpoints from TÜİK's current table, or both from the Ministry bulletin. One of each is not a comparison, it is a mix of two counting systems. Same definition: departing visitors with departing visitors, foreign arrivals with foreign arrivals, visitor income with visitor income rather than total revenue that includes transit passengers. If you are unsure which count a headline uses, read total visitors or foreign visitors: which number you are quoting.

Both endpoints dated: name the bulletin and its release date for each end, so a reader can reproduce the calculation. No derived rate across the revision break, and no year-to-date total built by adding quarters. TÜİK publishes quarters, not a running annual figure, so adding them produces a number nobody published.

2026 stays flat. Describe it as flat, and do not turn two quarters into an annual growth claim. Say nominal when you quote dollars over years, and add the per-visitor and nights columns when the point is about value rather than volume.

What can a national series tell a hotel, and what can only its own numbers tell it?#

The national series tells you the direction and the mix of the market; it cannot tell you how your property, your province or your season performed, because it is a national average across all departing visitors and every purpose of travel. Ministry accommodation statistics are a separate dataset, and they too average across every certified facility.

What it gives you: the direction of demand, the weight of source markets, average length of stay and average spend per visitor, all at national level and all dated. What it cannot give you: your province, your segment, your season, your channel mix or your rate position. A national average is a denominator, not a benchmark for your hotel.

The workable method is to compare your own year with your own year, then use the national line to explain the context in one sentence. Your history supplies the numerator, the official series supplies the background.

Read the definitions before you borrow a number. A figure built from all certified facilities, including small accommodation, is not a hotel figure, and an average over twelve months hides the season your property actually lives on. If you need geography, that is a separate exercise with its own coverage limits, not a column you can pull out of the national total; see province-level tourism data in Türkiye and its limits.

Keep the disclaimer with the number: public data, cited with its publisher and date, implies no relationship between the publisher and whoever quotes it.

What to do next#

Before you quote a year, open the current TÜİK table and the Ministry bulletin yourself, write the release date next to the figure and keep the two series apart. Then do the only comparison that decides anything for your property: this year against your own last year, with the national line beside it as context rather than as a target.