{ "schema": "garnet-schema-pr-packs/v1", "description": "Anonymized real-shape JSON-LD PR packs we've shipped to clients. Each pack is paste-ready — drop the `after` block into your , replace placeholders, remove the `before` block if present.", "packs": [ { "id": "001-organization-sitewide", "title": "Site-wide Organization JSON-LD", "priority": "high", "lane": "geo", "applies_to": "every page on a B2B site (typically in shared head template)", "summary": "Most consulting sites have no Organization JSON-LD at all. Adding it is the single highest-leverage schema move — engines use it to disambiguate the firm from look-alike domains and weight subsequent Service/Article markup against a known publisher entity.", "engine_impact": "ChatGPT and Gemini both lean on Organization JSON-LD for publisher disambiguation. Without it, multi-firm queries ('best ERP consultants') tend to cite Wikipedia or industry-list aggregators instead of source firms.", "before_state": "No JSON-LD or only a minimal {\"@type\":\"WebSite\"} blob", "after": "", "verification": "After deploy, validate at https://validator.schema.org. Then check Google Rich Results: https://search.google.com/test/rich-results" }, { "id": "002-service-offer-tier-pricing", "title": "Service + Offer with tier pricing on each lane/product page", "priority": "high", "lane": "geo", "applies_to": "every service/product landing page (e.g. /services/audit, /pricing/pro)", "summary": "Engines treat priced services with explicit Offer markup as buyer-actionable. Without UnitPriceSpecification, your page reads as a marketing page; with it, the page reads as a product the engine can recommend with a price.", "engine_impact": "Perplexity and Gemini both surface explicit monthly prices in their answer text when Service.offers.priceSpecification is present. Pages without it get summarized at the firm level instead.", "before_state": "Plain HTML with price visible in a hero card but no JSON-LD", "after": "", "verification": "Validate at https://validator.schema.org. The Offer array shows each tier independently; engines will quote a single tier in answers, so make tier names short and human (Pro/Scale/Enterprise rather than tier-1/tier-2)." }, { "id": "003-breadcrumblist-methodology", "title": "BreadcrumbList on every non-home page (especially methodology / how-it-works)", "priority": "high", "lane": "geo", "applies_to": "every page below the root — methodology, articles, case studies, contact, pricing", "summary": "BreadcrumbList is one of the cheapest wins. It costs nothing, requires no editorial effort, and gives engines a clean parent-child path through the site. Pages without it tend to get cited in isolation, losing the context that 'this is part of a larger methodology page set.'", "engine_impact": "Claude and Gemini both use breadcrumbs to reconstruct site hierarchy when summarizing. Without breadcrumbs, multi-page methodology series get summarized as disconnected pages — the engine doesn't surface 'see also' relationships well.", "before_state": "No breadcrumb markup of any kind", "after": "", "verification": "Each ListItem.item should match the page's canonical URL exactly. Validate with https://validator.schema.org. Per-page BreadcrumbList is server-rendered (not JS-injected) so engines that don't execute JS still see it." }, { "id": "004-faqpage-methodology", "title": "FAQPage JSON-LD on methodology / how-it-works pages", "priority": "medium", "lane": "geo", "applies_to": "methodology pages, how-it-works pages, FAQ pages — any page where you answer common buyer questions", "summary": "FAQPage markup is what surfaces in 'People Also Ask' style assistant answers. The question phrasing in the markup IS the prompt the engine matches against — write the questions in your buyer's voice, not your marketing voice.", "engine_impact": "ChatGPT and Perplexity both pull verbatim from FAQPage answers when the user's query closely matches a Question. This is the highest-conversion-rate citation form: the engine quotes you, attributes you, and links you.", "before_state": "FAQ exists as visible HTML accordion but no JSON-LD", "after": "", "verification": "Validate with https://validator.schema.org. Use 4-8 questions per page max — engines de-prioritize FAQPage blocks with 20+ items as keyword-stuffing." }, { "id": "005-article-with-author-bio", "title": "Article with Person author + publisher Organization on every content page", "priority": "medium", "lane": "geo", "applies_to": "every blog post / insight / case study / report", "summary": "Article markup with a concrete Person author (linked to a real about-page URL) and Organization publisher both lifts E-E-A-T signal. The author URL should resolve to a real bio page; engines follow it and use the bio as additional context.", "engine_impact": "Google-Extended (Gemini's training crawler) gives noticeably higher weight to Article pages with named Person authors than to anonymous content. Anonymous-author articles get summarized but not attributed by name in answers.", "before_state": "Article exists but no JSON-LD, or generic '@type:WebPage'", "after": "", "verification": "The `author.url` must resolve to a real bio page (404 hurts trust signal). The `image` URL should be ≥1200x630px (OG card size); smaller images get downranked in card-style answer surfaces." }, { "id": "006-howto-methodology", "title": "HowTo on every methodology / how-it-works page", "priority": "medium", "lane": "geo", "applies_to": "methodology pages, /how-it-works, /process — any page that documents a multi-step service delivery", "summary": "HowTo with HowToStep array is what engines surface in 'how does X work' queries. The step text gets quoted verbatim. Add totalTime, supply, and tool fields for completeness — they shift the page from 'description of a process' to 'attributable instructions'.", "engine_impact": "Gemini and ChatGPT both extract HowToStep arrays when answering process questions. Pages without HowTo get paraphrased; pages with HowTo get cited with the step name + a snippet.", "before_state": "Multi-step content visible as