Methodology
How SortaRich ranks — and what we don't do.
Every score on this site comes from public-domain institutional data — World Bank, OECD, WHO, IEP, Eurostat — combined with your own slider weights. The formula is visible. The data sources are named. The honest limitations are admitted below.
1. How rankings work
SortaRich uses a two-stage personalized ranking engine. There is no opaque composite formula behind a single "9.2/10" score — your ranking IS the slider weights you set, applied to normalized public-data dimension scores.
- Quiz answers → weight vector. Each dimension you mark as essential becomes a hard gate. Each dimension you weight matters becomes a soft preference.
- Hard-gate filter. Destinations that fail any active gate are dropped from the ranking entirely (e.g., LGBT-essential gate filters out criminalizing-tier countries).
- Soft-tiebreak rerank. Surviving destinations are ranked by a weighted sum:
score = Σ(weight_i × normalized_score_i), with adaptive-α adjustment for circumstance (segment, household, passport). - Anchor to your home city. All cost figures are PPP-adjusted to your home city — not a generic global average. Your $5,000/mo in Toronto buys differently than your $5,000/mo in Singapore.
- Confidence-decorated. Each destination card shows a confidence badge — see Section 3 below.
2. Per-dimension scoring
Each dimension below is normalized to 0–100 (higher = better). Sources are public-domain except where noted. The full source list lives in the engine code at app/lib/quiz-weight-calculator.ts.
| Dimension | Inputs | Sources | Refresh |
|---|---|---|---|
| Cost of living | PPP-adjusted price level + city rent indices + detailed prices | World Bank ICP 2021, Eurostat, Numbeo (validation only) | Annual + daily currency |
| Safety | GPI + WGI political stability + homicide rate + city-level perceived crime | IEP Global Peace Index 2025, World Bank WGI | Annual |
| Healthcare | UHC service coverage + physician density + life expectancy | WHO Global Health Observatory | Annual |
| Education (public) | PISA + school life expectancy + tertiary enrollment | OECD PISA 2022, World Bank | Triennial (PISA), annual otherwise |
| International schools | School-by-school tuition + curriculum + accreditation | International Schools Database scrape + manual verification | Quarterly per-city refresh |
| Mobility (visa) | Visa programs by passport-of-origin × destination × visa-type | SortaRich visa database (sourced from official immigration authorities) | Monthly |
| Economic | GDP per capita PPP + median wage + Gini | World Bank, ILO ILOSTAT | Annual |
| Domestic help | Childcare cost + housekeeping cost | SortaRich product-price database | Quarterly |
| LGBT friendliness | ILGA Rainbow Index + Equaldex public-policy + safety + same-sex marriage recognition | ILGA-Europe Rainbow Index 2025, Equaldex | Annual |
| Climate | Historical temperature, rainfall, climate comfort | Open-Meteo | Daily |
| English prevalence | EF EPI | EF English Proficiency Index 2024 | Annual |
| Happiness | World Happiness Report ladder score | WHR (Sustainable Development Solutions Network) | Annual |
3. Data confidence
Every city + country page shows a DataConfidenceBadge — a 4-tier indicator (excellent / strong / moderate / limited) that aggregates 12 weighted categories. The tier reflects how much of the destination's profile comes from source-backed data versus modeled estimates or country-level fallbacks.
City-level indices on cost, safety, healthcare, schools, prices. Most US/EU/East Asian capitals.
City-level indices on most dimensions; one or two country-level fallbacks. Most secondary cities + most of LATAM/SEA.
Mix of city-level + country-level fallback data. Smaller cities or countries with thinner subnational coverage.
Mostly country-level fallbacks. Reliable for ballpark, not for tactical decisions. Treat as exploratory.
4. Subnational coverage (the honest one)
No relocation site we've seen publishes this. SortaRich tracks 196 countries with 319 distinct spatial-data sources. Here is how that breaks down. The DataConfidenceBadge on each city page reflects this directly.
| Tier | Count | What it means | Example |
|---|---|---|---|
| Regional (genuine subnational) | 139 | City-level price data sourced from a region-level statistics agency, national CPI broken down by metro, or scraped per-city listings. | US (BLS regional CPI), Germany (Destatis Länder), Brazil (IBGE metropolitan) |
| National smear (honest fallback) | 54 | Only national CPI is available — we apply it uniformly to all cities in the country. Honest about the limitation. The DataConfidenceBadge marks affected city pages. | Bangladesh, Albania, Madagascar |
| Spatial-grid | 3 | Literal spatial-grid price data (price-point lookups by lat/long). | Costa Rica, Iceland, Singapore |
5. What we explicitly do not do
We don't use Numbeo as a primary signal
Numbeo is crowd-sourced and prone to small-sample bias. We use it only for validation against institutional sources, never as the headline cost figure.
We don't accept payment for ranking position
No advertiser, partner, or affiliate has ever influenced a country's rank. The engine has no concept of "advertiser" — it is a function of your quiz, your home city, and public data only.
We don't hide the formula
Your slider weights are the formula. SortaRich's ranking is `weighted_score = Σ(weight_i × normalized_score_i)` — completely visible. Competitors who report "Country X scored 9.2/10" with no exposed formula are running an opaque composite.
We don't pretend country-level data is city-level
When city data is missing, the DataConfidenceBadge tells you. We don't paint a country-level number across 50 cities and call it "city-level data for 50 cities." The 54 national-smear countries above are listed honestly.
6. Methodology FAQ
How does SortaRich rank countries?
SortaRich uses a transparent two-stage personalized ranking. Stage 1: hard gates from your quiz answers filter out destinations that fail your essential criteria (visa eligibility, safety floor, walkability requirement, etc.). Stage 2: surviving destinations are ranked by a weighted soft-tiebreak across the dimensions you weighted most heavily. The weights are visible to you in the slider UI — there is no opaque internal composite formula. Your ranking is anchored to your home city via real subnational PPP, not a generic global average.
What data sources does SortaRich use?
Public-domain institutional sources only — World Bank International Comparison Program (ICP) 2021 for purchasing power parity, OECD PISA 2022 for education, WHO Global Health Observatory for healthcare, IEP Global Peace Index for safety, EF English Proficiency Index for English, Eurostat for European cost data, ILGA Rainbow Index + Equaldex for LGBTQ+ rights, Pew Research for religious composition, Penn World Tables 11.0 for cross-validation. We deliberately do NOT use Numbeo as a primary signal because crowd-sourced opinions are not data.
Why is data confidence different from data accuracy?
Confidence measures how much of a destination's profile comes from source-backed data versus modeled estimates or country-level fallbacks. A city with city-level cost indices, city-level safety data, and source-backed school listings scores "excellent" (90%+). A city where we only have country-level fallbacks scores "limited" (< 45%). Accuracy is a different question — even our highest-confidence data has institutional measurement error. Confidence tells you how much we are reading the destination rather than guessing it.
What does subnational coverage actually mean?
For 139 of 196 countries SortaRich has genuine subnational price data — region-level CPI, city-level rent indices, etc. For 54 countries we apply national CPI uniformly across cities (we call this "national smear" — it is honest but less personalized). For 3 countries we have spatial-grid coverage. The methodology page below this explains the breakdown by country. We are the only relocation tool that publishes this admission — competitors who claim "city-level data for 380 cities" are mostly applying country-level data to those cities without saying so.
How often is the underlying data refreshed?
Tier-1 sources (World Bank ICP, IEP GPI, OECD PISA, Eurostat, WHO UHC) refresh annually on the source's own publication schedule — typically Q1-Q2 of each year. SortaRich's data-loading jobs run daily and pull whatever the latest published version is. Tier-2 sources (currency rates, weather data) refresh hourly. Per-page "Data updated [date]" timestamps reflect the most recent successful pull.
Does SortaRich accept payment for ranking position?
No. SortaRich does not accept payment for inclusion or ranking position. Affiliate revenue from optional paid services (insurance, banking partnerships) does not influence the ranking engine in any way — the engine has no concept of an affiliate signal. The ranking is purely a function of your quiz answers, your home city, and the underlying institutional data.
See the methodology applied to your situation.
See where you're sorta rich →Last reviewed: 2026-05-01. The ranking engine source lives at app/lib/weighted-audience-data.ts + app/lib/quiz-weight-calculator.ts and is open for inspection on the public repository.