ZenDawn

Methodology

How we collect, verify, and update retreat data. Why our rankings show what they show. What we exclude and why. Where commission enters and where it does not.

Last reviewed: 2026-05-04 · Directory size: 500 verified centers across 50 states · Pricing coverage: 100%

1. Data Sources

Every retreat listing in ZenDawn is enriched directly from the center's own published materials — primarily the center's official website, supplemented by their published program calendars, FAQ pages, and pricing pages. We do not pull data from third-party booking platforms or aggregators, because their representations of a center's pricing, schedule, and policies often lag behind the source.

For each listing we capture: name, address (street, city, state, lat/lng), website URL, phone, free-form description, traditions practiced, silence levels offered, retreat durations available, price minimum and maximum (USD), price model (fixed / sliding-scale / by-donation / contact-for-pricing), accommodation types, dietary options, beginner-friendliness, ADA accessibility, and teacher names where available.

2. Verification & lastVerified Cadence

Every listing carries a lastVerified date that surfaces on the listing page and feeds directly into the sitemap's <lastmod> per URL. State, city, tradition, duration, and budget hub pages inherit their lastmod from the most-recent lastVerified across the listings they contain.

Listings are re-verified on a rolling cadence. When a center's source page returns a 404, when their website changes substantively, or when a flag from a user or our own pipeline indicates a data conflict, that listing moves to the front of the re-verification queue. Listings that cannot be re-verified after multiple attempts are removed from the directory rather than allowed to stale.

3. Pricing Methodology

Pricing data is captured as a range (priceMin through priceMax, in USD) plus a priceModel field. The model field distinguishes:

  • Fixed — published price per person, per program
  • Sliding scale — tiered pricing based on financial capacity
  • By donation (dāna) — no fixed fee; suggested or freely-given contribution after the program
  • Contact — pricing not published publicly; user must inquire directly

Where a center publishes multiple program tiers (e.g., shared dorm vs. private room, weekend vs. 7-day), priceMin reflects the lowest published option and priceMax reflects the highest. Where a center publishes only "from $X", we record priceMin only and leave priceMax null. We do not infer pricing where none is published.

4. Exclusion Criteria

ZenDawn aims to be comprehensive within scope. The following are out of scope and are not added even when they appear in adjacent searches:

  • Hotels, resorts, and timeshares offering one-off "wellness weekends" without a structured retreat program
  • Single-teacher event listings (not centers); we list venues, not specific teacher offerings
  • Centers without a verifiable public website or whose claimed location cannot be confirmed
  • Centers with active legal or ethical action against them (we re-evaluate when the situation resolves)
  • Centers explicitly outside the United States (a US-only directory by design)

5. Comparison Page Generation

Head-to-head comparison pages at /retreats/compare/{a-slug}-vs-{b-slug}/ are auto-generated for retreat pairs that satisfy at least one of:

  • Same city (e.g., two retreats both in Sebastopol, California)
  • Same state and share at least one tradition (e.g., two California Vipassana centers)

Pair generation is bidirectional-deduped via sorted ID pairs — every pair appears exactly once. The current build produces 554 unique comparison pages from the 500 source listings. No editorial selection happens at the pair-generation step; the algorithm is deterministic and reproducible from the source data.

6. Ranking & Sort Order

Listings within a state, city, tradition, duration, budget, or audience hub appear in the order they are stored in retreats.json, with no commercial ranking layer. There is no "featured" placement, no "sponsored" tier, and no commission-driven ordering. Filter pages may apply data-driven sub-ordering (e.g., the donation-based hub surfaces explicitly donation-model centers first), but those rules are described inline on each page when applied.

Comparison pages present the two retreats side-by-side in a deterministic A/B order based on slug sort — not on price, not on commission, not on editorial preference.

7. Conflict-of-Interest Policy

Some "Visit" buttons on individual listing pages route through commercial-booking-platform affiliate links. When a user books through one of those links, ZenDawn may receive a commission at no additional cost to the user. As of the last review date above:

  • Roughly 24 of 500 listings (~4.8%) have any affiliate path available at all — the remaining 476 link directly to the center's own website
  • The presence or absence of an affiliate link does not influence whether a center appears in the directory
  • The presence or absence of an affiliate link does not influence sort order, comparison pairing, or hub placement
  • Where an affiliate link exists, it carries rel="sponsored noopener" per Google's published guidance and per FTC disclosure norms
  • Where no affiliate link exists, the "Visit" button routes directly to the center's own website with rel="noopener"

8. Affiliate Programs (Active & Pending)

Active programs as of the last review:

  • BookRetreats — 6 listings currently routed through affiliate links

Pending applications:

  • Tripaneer / BookYogaRetreats — estimated coverage of ~19 additional listings on approval
  • Mindfulness.com — for post-retreat integration funnel (cornerstone-content surface)

We will update this list whenever a new program goes live. Programs are evaluated on whether they would be recommended without commission — a program that fails that test is dropped, regardless of payout.

9. Data Corrections

If you believe a listing on ZenDawn is inaccurate — pricing has changed, a tradition tag is wrong, the center has closed, the website has moved — please let us know. Corrections are processed in batches and the affected listing's lastVerified date is updated upon re-verification.

10. Open Questions We Are Still Resolving

  • Audience tagging beyond beginner / experienced — we support beginner-friendly and experienced tags powered by the existing beginnerFriendly field. Specialized audience filters (families, couples, women-only, LGBTQ+ affirming) require a per-listing audit of explicit programming and welcoming statements. Until that audit is complete, those filters are not surfaced — we prefer no listing to a misleading one.
  • Float tank, sound healing, and breathwork directories — the URL templates exist; small public-data foundations have been seeded as proofs of concept. Larger coverage requires either partnerships with existing directories or a structured manual audit.
  • International expansion — out of scope at this time; the current directory is US-only by design.

Methodology last reviewed: 2026-05-04. We update this page whenever a substantive change is made to data sources, verification cadence, exclusion criteria, ranking rules, or conflict-of-interest disclosures.