Should you really limit your JSON-LD properties to just what's shown in rich snippets?

What does this recommendation from Google really mean?

Google tells us to limit our JSON-LD properties to what can realistically appear in rich snippets. For recipes, this includes visual elements: image, preparation time, calories, overall rating. The underlying idea is simple: why structure data that the engine will never display?

This position contrasts with a maximalist approach where some SEOs add all possible properties from schema.org/Recipe. Cooking time, resting time, marinating, necessary equipment, precise yield, multiple categories, cooking keywords... Everything is included. Google suggests that this overload is unnecessary.

Why does this statement raise questions?

The issue lies in the ambiguity of the term "compatible". Compatible with what exactly? With the current display of rich snippets, or with a future version Google might deploy? The SERPs are constantly evolving, and what is not displayed today could be shown tomorrow.

Furthermore, Google does not say that extra properties are penalizing. It simply states to ensure they are "compatible". This nuance leaves strategic ambiguity. Is it a recommendation for optimization or a warning against structured data spamming?

Which properties are actually displayed in rich snippets?

For recipes, the standard SERP display includes: image, title, aggregate rating, number of reviews, total time, calories. Sometimes, you also see preparation and cooking times separated, but this is less systematic. Yield (number of servings) rarely appears.

Properties such as recipeCategory, recipeCuisine, keywords, tool, suitableForDiet almost never display in traditional organic results. They might serve internal ranking or future filters, but no public data confirms this.

• Essential properties: name, image, author, datePublished, description, aggregateRating, recipeYield, prepTime, cookTime, totalTime, recipeIngredient, recipeInstructions, nutrition (calories)
• Properties displayed in rich snippets: image, aggregateRating, totalTime (or prepTime + cookTime), nutrition.calories
• Rarely utilized properties: recipeCategory, recipeCuisine, keywords, tool, suitableForDiet, video (unless in a specific context)
• Required properties for validation: name, image, author, datePublished are mandatory even if not all are displayed
• Gray area: video sometimes becomes a separate carousel, nutrition beyond calories does not appear, detailed steps (HowToStep) are validated but rarely fully displayed

Is this recommendation consistent with observed practices in the field?

Not really. Sites that rank highest on recipe queries often use rich implementations with far more than just the minimum displayable. They structure ingredients in detail, steps in HowToStep, add video, recipeCategory, recipeCuisine. Is it correlation or causation? Impossible to say with certainty. [To be verified]

Google has never publicly confirmed that non-displayed properties contribute to ranking. However, several A/B tests conducted by agencies show that adding additional properties sometimes coincides with improved visibility. The causal link remains unclear, but the correlation exists.

What nuances should be added to this directive?

The first nuance: “compatible” does not mean “only”. Google does not say to exclude any non-displayed properties. It says to ensure what you add is compatible with what can display. In other words, do not invent fictitious or poorly structured properties.

The second nuance: additional properties may serve other Google products. Google Assistant, Google Discover, Google Lens potentially use data that classic Search does not display. Limiting your JSON-LD to just the SERP display ignores these channels.

In which situations does this rule not apply or become counterproductive?

If your content strategy relies on semantic richness for voice assistants or third-party aggregators, then limiting your properties is a mistake. Structured data feeds much more than rich snippets: it fuels knowledge graphs, search suggestions, and advanced filters.

Another case: niche sites with highly specialized recipes (specific diets, regional cuisines) benefit from finely structuring suitableForDiet, recipeCuisine, recipeCategory. Even if these properties do not display today, they prepare the ground for future filters.

What concrete actions should you take with your current Recipe tags?

First step: audit your JSON-LD implementations to identify the properties present. Use Google's Rich Results Test tool, then compare what is validated against what actually displays in the SERP for your recipe pages. Note any discrepancies.

Second step: prioritize essential properties. Ensure that name, image, author, datePublished, aggregateRating, totalTime, recipeIngredient, recipeInstructions, nutrition.calories are present and correctly formatted. These are the pillars of the rich snippet. Without them, there is no rich display.

What mistakes should you avoid during implementation?

Classic mistake: overloading with redundant or poorly formatted properties. For example, adding prepTime, cookTime, AND totalTime when totalTime should be the sum of the two. Google can get confused or ignore certain values. Stay consistent.

Another pitfall: using non-standardized or outdated properties. Schema.org evolves, and some properties are deprecated. Regularly check the official documentation. A poorly named or typed property simply won’t be read, and you waste crawl time for nothing.

How can you check if your implementation is optimal?

Test your URLs in Google Search Console's Rich Results Test tool. Look at warnings, not just errors. A warning on a missing property could signal an opportunity for improvement, not necessarily an obligation.

Then compare the actual display in the SERPs. Conduct incognito searches on your target queries and see what appears. If a property you structured never displays, there are two options: either it is unnecessary for display (but useful elsewhere), or it is poorly formatted.

• Audit all recipe pages with the Rich Results Test tool
• Check for the presence and validity of essential properties: high-resolution image (minimum 1200px), aggregateRating with ratingValue and ratingCount, totalTime in ISO 8601 format
• Clean up non-standardized or redundant properties that generate warnings
• Test the real display in SERPs for at least 10 representative recipes
• Document additional properties that you retain (recipeCategory, recipeCuisine, video) and justify their strategic presence
• Set up monthly monitoring of rich snippets to detect any display loss