About the Index

A public benchmark for social preview reliability

We analyze HTML metadata quality across a stable ecommerce cohort. Cohort metrics are public; domain lookup is diagnostic only.

How the index works

Analyze a Domain

Subdomains and full URLs are normalized automatically.

Reference

Frequently Asked Questions

This page clarifies benchmark scope, methodology, eligibility rules, and interpretation boundaries.

Snapshot full_20260227-002835Feb 27, 2026

Scope & Inclusion

Why are only 647 domains included?

Public benchmark aggregates include only benchmark-eligible retail catalog domains for the snapshot window. For the latest public snapshot (full_20260227-002835, Feb 27, 2026), that yields 647 included domains after eligibility gates are applied.

Why is my domain not included?

A domain may be absent from benchmark aggregates if it is outside the current cohort, does not meet retail catalog eligibility criteria, or does not pass minimum coverage and URL-check thresholds during the snapshot window.

Why are country domains treated separately?

Registrable domains are the unit of analysis. Country domains are measured independently because architecture, catalog structure, and metadata behavior can differ materially across market-specific properties.

Measurement Model

Do you execute JavaScript?

No. Measurement is HTML-only. The crawler evaluates metadata available in server-rendered HTML responses and metadata probes.

What does the score represent?

The score summarizes structural social preview reliability signals on a 0–100 scale. Higher values indicate stronger structural reliability for link preview rendering.

What counts as Unreliable?

Unreliable represents a high likelihood of preview issues from structural signals such as missing or inaccessible image metadata and other severe metadata inconsistencies.

Why might a domain have a score but not influence the benchmark?

Domain-level lookup and benchmark aggregation are intentionally separated. A domain can have a measured score while remaining excluded from benchmark aggregates if eligibility criteria are not satisfied for the published snapshot.

Interpretation & Usage

Is this a ranking?

No. The public benchmark publishes aggregated distributions and methodological context, not domain leaderboards or ranking tables.

Does a low score mean a site is broken?

Not necessarily. A low score indicates elevated structural risk for degraded or broken social previews. It should be interpreted as reliability risk, not a blanket outage statement.

How often is the index updated?

The index updates on snapshot publication cadence. Each public release includes a bounded data window, publish date, and benchmark sample size.

Does blocking crawlers affect the benchmark?

Yes, blocking and rate limiting can affect measured crawl accessibility. These effects are reported as aggregate transport metrics and do not imply user-facing downtime.

How can I improve my score?

Prioritize structural metadata quality: ensure preview image tags are present and accessible, reduce canonical/OG inconsistencies, and improve product-page metadata consistency at scale.