1. Collect farm records
Farm records can come from farmer submissions, farm websites, public directories, official state or local agriculture resources, product-specific source lists, and manual research. The goal is to identify working farms and direct-to-buyer buying paths, not to mirror every generic business directory row.
2. Normalize location and product signals
Records are normalized into public fields such as farm name, city, state, coordinates, product categories, fulfillment options, website, phone, email, source hints, and last-reviewed or last-verified dates when available.
Category and local pages are generated only when there is enough supply to support the search intent. Thin, stale, duplicate, missing-state, non-US, or source-suspect farm records should not become search entry points.
3. Gate pages before sitemap exposure
The static page system decides whether a page belongs in the sitemap. Farm-profile pages need enough useful public detail. State, city, and category pages need supply-backed coverage. Utility, embed, duplicate, or intentionally skipped pages are kept out of the indexable sitemap surface.
4. Show useful facts, not hidden claims
Structured data should match visible page content. If a page says a farm sells a product, offers delivery, has a source, was reviewed, or is claimed, that fact should be visible or traceable on the page rather than hidden only in JSON-LD.
5. Keep correction paths visible
Farm data changes. Public pages should make it easy for farms and buyers to claim listings, report stale details, add missing products, and point to better sources. Corrections are part of the data-quality loop, not an afterthought.