How we curate the LoongBuy spreadsheet.

A deliberate catalogue snapshot

The published site reads from a saved local product dataset. It does not request a fresh API response whenever the preview server starts or the static pages are rebuilt. That separation keeps ordinary builds predictable and prevents product order, page totals and searchable copy from changing without review.

What makes an entry useful enough to show

An entry needs enough information to support comparison: a recognisable title, a category, at least one usable image and an indicative price where the source provides one. We favour products that are visually clear, relevant to an active category and useful beside similar alternatives.

• Clear enough title and image for product identification
• Category mapping that supports like-for-like comparison
• Indicative price data when the source provides it
• QC-photo availability recorded when present
• Balanced coverage across winter wear, footwear, clothing and accessories

Inclusion is not an endorsement. It does not establish authenticity, product safety, seller reliability, stock or shipping performance.

Focused category pages, not one endless feed

The current snapshot is condensed into 15 active visitor-facing category groups. This makes related products easier to compare and gives each search intent a focused page. Category names are normalised for navigation; they do not claim that every source record used exactly the same label.

QC-backed entries receive priority, not automatic approval

The orange QC badge is generated from the saved record’s QC-photo count. When a category contains QC-backed entries, they are useful early comparison candidates because more visual evidence may be available. We do not convert that count into a quality score, and we do not claim that every photo set has been personally approved.

Visitors should inspect overall shape, materials, stitching, hardware, labels, placement and measurements. If the evidence is incomplete or inconsistent, pause and request more information rather than treating the badge as a green light.

Displayed prices are orientation points

The site shows an indicative USD estimate from the saved product record when one is available. It is designed for rough comparisons inside a category, not for calculating a final order total. Currency movement, domestic delivery, service fees, international shipping, insurance, taxes and listing changes may all affect what a buyer ultimately pays.

Stable ordering with useful priorities

The catalogue is not re-randomised on each page load. Home-page selections can be deliberately adjusted to replace weak imagery or improve the mix of categories, while QC-backed records are surfaced as useful evidence-rich candidates. Those editorial choices are about clarity and coverage, not hidden sales claims.

Fetch, review, rebuild, publish

1. Fetch intentionallyNew catalogue data is requested only when a refresh is planned.
2. Review the snapshotCounts, categories, image quality and obvious weak entries are checked before publication.
3. Build static pagesThe approved snapshot becomes stable crawlable HTML, category pages and structured data.
4. Publish one versionThe refreshed site and its visible update record move together.

This workflow protects consistency for visitors and search engines while still allowing controlled catalogue improvement.

What this method cannot prove

No catalogue snapshot can guarantee authenticity, legality, suitability, stock, seller conduct, product condition, delivery speed, customs outcomes or final price. External content may change after publication. Always verify the current listing, applicable rules and the evidence relevant to your decision.