Why Numbeo Is Wrong About Thailand (And How We Count Instead)

Why Numbeo Is Wrong About Thailand (And How We Count Instead)

Numbeo's entire Bangkok cost-of-living page is built on 198 contributors submitting 1,365 prices over the prior 12 months, as of its 8 July 2026 update. Chiang Mai's page rests on just 54 contributors and 728 entries. Those aren't bad-faith numbers, but they are thin, self-selected, and prone to running stale, which is exactly why Numbeo's Bangkok rent average sits well above what real listings show and why we build our own figures from dated, cross-checked sources instead of quoting a crowd average as fact.

This page exists because every cost page on this site cites a number, and readers deserve to know where that number came from and why it might differ from what they see on Numbeo, Nomad List, or a forum thread. Below is the specific breakdown, followed by our methodology in plain terms.

Quick answer: the four Numbeo problems

• Self-selection: contributors are expats and nomads who feel strongly enough to submit a price, not a random sample of residents — this skews toward urban, prime-district, expat-facing prices.
• Thin samples: Bangkok's page runs on 198 contributors; Chiang Mai's on 54. A handful of expensive-condo submissions can move a city average meaningfully.
• Stale entries: Numbeo's own policy allows data up to 24 months old to remain visible in lower-contributor categories, quietly aging inside a "current" average.
• Rent skew high: Numbeo's Bangkok city-centre one-bedroom average (~฿22,759, ~$690) sits above the range on-the-ground listing sites (Hipflat, DDproperty, FazWaz) show for comparable units, because self-reported "what I pay" data leans toward newer towers and prime addresses.

The core problem: who actually submits a price

Numbeo is crowd-sourced. Anyone can submit what they say they pay for rent, a beer, or a taxi ride, and the site's own methodology page confirms manually gathered data (from government and retailer sources) is entered only twice a year and weighted three times higher than user submissions — meaning the bulk of what you see for day-to-day items like rent and groceries is still user-submitted, unweighted, and unaudited.

The people who bother to submit a Bangkok rent figure skew heavily toward expats and digital nomads living in condo towers along the BTS line, because that is the population reading English-language cost-of-living sites in the first place. A Thai family renting a walk-up near Huai Khwang for ฿8,000 a month almost never shows up in the sample. That is self-selection bias, not fraud, but it pulls the average toward the prime, English-speaking, high-rise slice of the rental market every time.

Thin samples, especially outside Bangkok

The self-selection problem gets worse the smaller the city. Numbeo discloses contributor counts directly on each city page, and the gap between Bangkok and everywhere else is stark.

Source: Numbeo Cost of Living pages, Bangkok accessed 2026-07-08 and Chiang Mai accessed 2026-07-02 · verified July 2026 · ฿33/USD.

Fifty-four contributors is not enough to describe a city of over a million people with any precision. A single active expat forum thread pointing new arrivals toward Nimman's priciest towers can shift Chiang Mai's entire average, and there is no way to see that from the headline number alone. Smaller cities like Hua Hin or Koh Samui, which don't even get their own dedicated Numbeo page in many cases, are worse off still.

Numbeo's rent number runs high, and here's why

Numbeo's own city-centre one-bedroom average for Bangkok (~฿22,759) lands above the ฿15,000–35,000 range that direct comparison against listing platforms such as Hipflat and DDproperty suggests for a comparable central one-bed in 2026. Three mechanics explain the gap:

1. Asking price vs. paid price. Listing sites show what landlords ask; Numbeo submissions are self-reported "what I pay," but expats disproportionately report their current lease rather than what they negotiated it down to, and negotiated discounts of 10–20% off the asking rent are common in Bangkok.
2. Prime-district overrepresentation. Submitters cluster in Sukhumvit-adjacent, BTS-served condo towers because that's where most English-speaking contributors live, pulling the "city centre" bucket toward the expensive end of that bucket.
3. No unit-size normalization. Numbeo's category is "1 bedroom apartment in city centre" with no consistent floor-area control, so a 65 sqm unit and a 35 sqm unit get averaged together, and newer, larger, amenity-heavy stock skews the mean up.

None of this makes Numbeo useless. It's a reasonable directional signal for "is this city more or less expensive than that one," which is why our Bangkok cost of living and Chiang Mai cost of living pages still cite it alongside other sources. The problem is treating a single Numbeo figure as a precise, current answer to "what will I pay."

Stale data hides inside the average

Numbeo's own archival policy, stated on its methodology page, sets a sliding threshold: a city needs more than 153 contributors in a trailing three-month window before data older than three months gets archived out, and more than 123 contributors in six months before six-month-old data drops. Below those thresholds, Numbeo can keep showing entries up to 24 months old as part of the "current" average.

Bangkok, at 198 contributors over 12 months, sits close to that lower threshold rather than comfortably above it, and most other Thai cities sit well below it. That means a chunk of what looks like a live July 2026 figure may actually be a rent someone reported in 2024, quietly averaged in alongside this month's numbers with no visible timestamp on the individual data point.

How we count instead

We don't run our own live rent survey, and we're upfront about that. What we do differently is treat every figure as a dated claim with a named source rather than a single crowd-sourced average presented as fact.

• We cross-check against listing sites, not just aggregator averages. Hipflat, DDproperty, and FazWaz show live asking prices for specific units, which we use to sanity-check any crowd-sourced average before publishing a range.
• We date every table and note the exact access date, so a number from July 2026 doesn't quietly stand in for December 2026 six months later.
• We hedge when sources disagree. If SmartLifeThailand's city-band figure and Numbeo's average sit ฿10,000 apart, we say so and give a range rather than picking whichever number is more convenient.
• We flag institutional prices as indicative, not fixed — this applies doubly to hospital pricing, where transparency has eroded since 2023 and readers should always get a written quote before treatment.
• We separate visa-money figures from cost-of-living figures. The foreign income tax status tracker and best health insurance for expats pages are sourced from regulatory filings and insurer rate cards, not crowd submissions, because those numbers carry real financial consequences if wrong.
• We publish ranges, not false precision. A single mean like "฿22,759" implies a precision the underlying sample can't support; we'd rather say "฿15,000–35,000 central, more in prime districts" and let you place yourself in that range.

If you want a number tailored to your own lifestyle rather than a city average, our cost-of-living calculator runs the line-item budget for your city, household size, and comfort tier, built from the same cross-checked sourcing described above, not a single crowd-sourced mean. For a full city-by-city view built this way, see the Thailand cost of living overview.

Frequently Asked Questions

Is Numbeo accurate for Thailand?

Directionally, yes; precisely, no. Numbeo is useful for ranking cities against each other (Chiang Mai cheaper than Bangkok, Phuket pricier than both) but its Bangkok page runs on only 198 contributors and can display rent and price entries up to 24 months old in lower-contributor categories. Treat it as a rough signal, not a live quote. See our own take at is Numbeo accurate for Thailand.

Why is Numbeo's Bangkok rent higher than what I see on listing sites?

Because Numbeo's contributors skew toward expats in prime, BTS-adjacent condo towers who self-report their current lease rather than a negotiated rate, and the "city centre" category isn't normalized for unit size. Listing sites like Hipflat and DDproperty show live asking prices across a wider mix of buildings and often land ฿5,000–10,000 below Numbeo's city-centre average for a comparable one-bed.

Does Numbeo update its Thailand data often?

Manually sourced data (from retailers and government) is refreshed twice a year, per Numbeo's own methodology page. User-submitted data, which makes up most day-to-day categories like rent and food, updates continuously but isn't purged until a city crosses Numbeo's contributor thresholds — below those thresholds, some entries can stay in the average for up to 24 months.

Should I ignore Numbeo completely?

No. It's a fine starting point for comparing two cities' relative cost, and its currency conversion and salary data are genuinely useful. Just don't budget your move around its exact rent figure — cross-check against current listings and a lived-budget source like ours before you sign a lease.

Where do Thailand Ultimate's numbers come from?

A mix of dated crowd-sourced data (Numbeo, Nomad List) cross-checked against listing sites (Hipflat, DDproperty, FazWaz), official tariff and fee schedules (utilities, hospitals, schools), and insurer or government rate cards for money-critical pages like tax and visas. Every table on this site carries a source line with an access date so you can judge its freshness yourself.

The bottom line

Numbeo isn't lying to you, it's just showing you an average built from a small, self-selected, occasionally stale sample, and presenting it with more confidence than the underlying data supports. Use it to compare cities, not to plan a specific budget. Every number on this site instead comes with a date, a named source, and a range wide enough to be honest about what a crowd-sourced average can't tell you.