Product & Offer Schema

This is a focused evaluation of structured data on pages representing purchasable products or services. Unlike schema coverage ratio (which measures any schema across all pages), this criterion zeros in on commerce-relevant pages and benchmarks them against the Product, Offer, AggregateOffer, and AggregateRating specifications.

We identify commerce pages through multiple signals: URL patterns (/product/, /shop/, /pricing/, /services/), content indicators (pricing info, "Add to Cart" buttons, availability status), and existing schema types. Each page is evaluated against Google Merchant Center's structured data requirements for Product rich results -the most stringent publicly documented standard.

The primary metric: "product schema completeness" -the average number of recommended Product properties present divided by total recommended, across all commerce pages. Google recommends 13 properties: name, image, description, sku, brand, offers (with price, priceCurrency, availability, url), aggregateRating (ratingValue, reviewCount), and review. We check each individually.

A secondary metric, "offer accuracy," validates that prices are numeric, currency codes are valid ISO 4217, availability uses Schema.org ItemAvailability enums (InStock, OutOfStock, PreOrder), and that structured data matches visible page content. A Product schema claiming $29.99 when the visible price is $39.99? That's a critical accuracy violation we flag immediately.

AI shopping assistants are one of the fastest-growing AI use cases. ChatGPT, Perplexity, and Google's AI Shopping experience all pull product information from structured data to generate comparison tables, price lists, and recommendations. Without Product schema, your products are invisible to these interfaces.

The coverage gap has direct revenue implications. When someone asks ChatGPT "What are the best live chat tools under $50/month?" the AI constructs a comparison by pulling Product and Offer schema from indexed sites. Products without schema aren't included -regardless of how well they match the query. We've seen this play out across the live chat vertical (Tidio: 63, LiveChat: 59, HelpSquad: 47/42): sites with complete Product schema appear in AI comparisons, sites without are systematically excluded.

For e-commerce sites with hundreds of products, coverage has to be template-level, not page-level. If your product page template includes Product schema, every product gets coverage automatically. If schema is manually added per product, coverage inevitably decays as new products are added without it. We measure template-level consistency by grouping commerce pages by URL pattern and checking whether coverage is uniform.

AggregateRating is high-value for AI citations. AI engines use star ratings and review counts as trust signals. Products with AggregateRating are more likely to be recommended because the AI can present verifiable social proof ("4.7 stars from 2,341 reviews"). Products without ratings data look less trustworthy in AI-generated recommendations.

We start by identifying all commerce pages. The detection algorithm uses weighted scoring: URL patterns with commerce keywords (+2), visible price elements (+3), "Add to Cart" or "Buy Now" buttons (+3), existing Product or Offer schema (+5), and product-like content structure (image + title + description + price layout, +2). Pages scoring above threshold (5 points) are classified as commerce pages.

Each commerce page gets its JSON-LD extracted and validated against a 13-property checklist organized into three tiers.

Tier 1 (required for basic eligibility): name, image (valid URL to an actual image), description (minimum 50 characters), and at least one Offer with price and priceCurrency. Missing any Tier 1 property = critical flag.

Tier 2 (required for rich results): sku or gtin, brand (Brand type with name), availability (valid ItemAvailability enum), and url (canonical product URL). Missing Tier 2 means basic indexing works but rich result opportunities are lost.

Tier 3 (recommended for AI citation): aggregateRating with ratingValue and reviewCount, at least one review with author and reviewBody, offers with priceValidUntil and seller. Tier 3 dramatically increases AI citation likelihood but isn't strictly required.

We cross-validate structured data against visible page content. For each Product schema, we check that name matches visible title (90% fuzzy match tolerance), price matches within 1%, and availability matches visible stock indicators. Mismatches are accuracy violations -search engines penalize sites with deceptive structured data.

For product variants (sizes, colors, configurations), we check whether variant-level Offers exist with distinct prices and availability per variant, or if the schema only represents the base product.

Product schema scoring applies only to sites with identified commerce pages. Sites without commerce pages receive "not applicable" rather than a low score.

1. Product schema coverage (3 points): - 90%+ of commerce pages have Product schema with all Tier 1 properties: 3/3 points - 70-89%: 2/3 points - 50-69%: 1/3 points - Below 50%: 0/3 points

2. Schema completeness (3 points): - Average 11+ properties per Product schema (out of 13): 3/3 points - 8-10 properties: 2/3 points - 5-7 properties: 1/3 points - Fewer than 5 (minimal schema): 0/3 points

3. Data accuracy (2 points): - All schemas match visible page content within tolerance: 2/2 points - Less than 5% mismatch: 1.5/2 points - 5-15% mismatch: 1/2 points - More than 15% or critical accuracy violations: 0/2 points

4. AggregateRating and Reviews (2 points): - 80%+ of Product schemas include AggregateRating: 2/2 points - 50-79%: 1.5/2 points - 20-49%: 1/2 points - Below 20% or no rating data: 0/2 points

Deductions: - -2 points if any schema contains fabricated data (fake reviews, inflated ratings) - -1 point if prices use the wrong currency code for the site's primary market - -0.5 points if variants exist but schema only represents the base product

E-commerce platforms with native schema support (Shopify with JSON-LD plugin, WooCommerce with schema plugins) typically score 5-8. Custom-built sites without structured data tooling score 0-3.