Product & Data Changelog
Material changes are recorded here so users can see when the evidence, methodology, or public explanation changes.
September 23, 2026 — Privacy-minimal site analytics
Added first-party aggregate public-page, landing, calculator-open, and successful-result events. A random per-tab value is transformed into a keyed one-way value before storage; it is not persistent across browser sessions. Reports use only coarse acquisition classes, exclude signed release and regression testing, suppress groups below five, and retain no IP address, user-agent, referrer URL/domain, cookie, fingerprint, or personal inputs. Wage data, country mappings, formulas, and methodology are unchanged.
August 21, 2026 — SEO foundation release
Added ten evidence-rich category hubs, retained twelve V2-native core occupation pages, and added six software-developer country analyses only where reviewed Certified Local Primary records support unique official-source content. Rebuilt the sitemap as a clean canonical allowlist and added permanent redirects for legacy .html and /pages/... paths. No alias landing pages or currency-conversion country×occupation pages were created.
August 21, 2026 — Methodology transparency and versioned FX release
Publication candidate TR-DATA-2026-08-21.1 formalizes the International Reference taxonomy, verifies 21 May 2025 BLS percentile records, correctly labels the separate real-estate model anchor, quarantines legacy country metadata, removes duplicated precomputed reference fields, and introduces fail-closed U.S. Treasury reporting-rate snapshot TR-FX-USTREAS-2026-06-30-r1. Methodology remains TR-METH-2.0.0; the provisional ×1.75 coefficient is unchanged.
August 20, 2026 — Privacy-minimal analytics baseline
Added first-party aggregate calculator events for page views, supported country and occupation selections, successful calculations, fixed validation-error codes, benchmark position, and source-detail expansions. Events contain no raw search text, personal financial inputs, calculated amounts, cookies, user/session IDs, referrers, or stored IP/user-agent fields. Source datapoints expire after three months; reports use sampling-aware counts and combine dimension cells below five events.
August 20, 2026 — Seven-country V2 and site consistency release
- Made the 98-core, 1,794-alias V2 calculator the public
/calculator experience. - Added separate Accountant, Bookkeeper, Lawyer / Legal Consultant, and Paralegal / Legal Assistant cores.
- Certified local occupational evidence for the United States, United Kingdom, Canada, Netherlands, Australia, Germany, and Sweden.
- Preserved official hourly, weekly, monthly, and annual units; suppression and lower-bound qualifiers remain protected.
- Removed the previous experience, client-market, purchasing-power, and premium-target adjustments from the live model. The current modeled planning benchmark applies the disclosed ×1.75 coefficient once.
- Rebuilt public skill guides and worked examples around current sources, mappings, limitations, and reproducible calculations.
August 12, 2026 — V2 candidate certification
Certified the occupation split, alias migration, U.S. evidence model, UK/Canada/Netherlands local mappings, international-reference boundaries, and source disclosures before production rollout.