Scope: This is an implementation reference. Validate claims against current Google, vendor, and platform documentation before making a release or policy decision.
The citation pipeline
A page can be technically crawlable and still be a weak citation. Answer systems need a passage that answers the query, a clear entity-to-claim relationship, and enough source context to attribute the statement. Dense navigation, vague introductions, and unsupported claims make extraction harder.
- Discovery: the URL must be reachable and eligible for crawling.
- Retrieval: a passage must contain query-aligned terms and a complete answer.
- Attribution: author, organisation, date, and source identity must be unambiguous.
- Answer composition: the passage must support a claim without misleading truncation.
Write citation-ready passages
Put the answer first, then qualification and evidence. One heading should introduce one question or task. Avoid making the only answer depend on five paragraphs of pronouns and marketing language.
<h2>What is INP?</h2>
<p>Interaction to Next Paint (INP) measures the latency of user interactions
across a page. A good INP is 200 milliseconds or less at the 75th percentile.</p>
<p>Measure it with field data and investigate long main-thread tasks when it fails.</p>This format is useful for users, classic search snippets, and passage retrieval. It is not a guarantee of citation; it is a reduction in ambiguity.
Entity hierarchy and evidence
Use structured data to describe visible facts, not to manufacture authority. Link the article to its author and publisher, expose dates, and keep the same names across the page, organisation schema, author page, and external references.
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "How to measure INP",
"author": { "@type": "Organization", "@id": "https://example.com/#org" },
"publisher": { "@id": "https://example.com/#org" },
"dateModified": "2026-09-12",
"mainEntityOfPage": "https://example.com/knowledge/inp"
}