Skip to main content

Public Is Not the Same as Indexed: SEO Lessons From 1,500 Recipe Pages

How a static recipe archive separates public, indexable, and promoted pages—and what actually improved crawl quality.

Updated 4 min readMarkdown
Public Is Not the Same as Indexed: SEO Lessons From 1,500 Recipe Pages

The Fancy Fox recipe archive grew to more than 1,500 static pages before search discovery became the project's most important technical question.

Our first instinct was too binary: either put a page in the sitemap or hide it from search. That confused three different controls.

Three states, not two

Public means a person can open the URL and use the page.

Indexable means the page does not prohibit a search engine from indexing it.

Promoted means we actively surface the page through the sitemap, homepage, important collections, and external distribution.

A page can be public and indexable without appearing in the sitemap. A sitemap is a discovery and prioritization signal, not a complete inventory requirement and not a guarantee of indexing.

That distinction let us keep the long-tail archive available while giving the best pages a clearer path.

Our 50-URL sitemap

The current sitemap deliberately contains 50 URLs:

  • the homepage;

  • 10 cuisine collections;

  • 5 category collections;

  • 4 diet collections;

  • 30 featured recipes.

The featured recipes are selected from engagement on Fancy Fox-owned social posts. A page such as crispy salmon with charred zucchini, roasted fennel, and basil-pistachio crema receives homepage placement and strong internal links. The rest of the archive remains reachable through collections, ingredients, and search.

What stayed indexable

We did not add noindex to every recipe outside the sitemap. Doing that would prevent a search engine from selecting a useful long-tail page it discovers through internal or external links.

Instead, those recipes remain self-canonical and indexable. We reserve noindex for pages that should genuinely stay out of search, not merely pages that are absent from today's promoted set.

Collection pages need real identity

Taxonomy pages can become thin permutations quickly. We reduced and cleaned the public taxonomy, then gave the important collection pages:

  • a descriptive title and meta description;

  • a self-canonical URL;

  • a visible recipe list;

  • collection-specific imagery;

  • crawlable links to the recipes;

  • valid ItemList structured data matching the visible entries.

For example, the Italian recipe collection is a real browsing page, not a keyword landing page with an empty template.

Structured data has to match the page

Recipe JSON-LD is valuable only when the underlying values are valid and visible content tells the same story. A single model placeholder such as "none" in an ingredient preparation field can make a generated page look careless even if the JSON parses.

We added validation before rendering and treated structured-data tests as build tests. The checks cover required recipe fields, canonical URLs, taxonomy lists, and sitemap policy.

Performance is part of crawl quality and human quality

The public pages are static, images have optimized WebP versions, and below-fold media is lazy-loaded. Those changes help mobile visitors first. They also make the site easier for crawlers to fetch consistently.

The biggest visual asset remains the recipe image, so optimization cannot mean removing it. The correct trade is a properly sized, compressed image with stable dimensions and a fast first render.

The remaining constraint is authority

Technical correctness makes a page eligible. It does not make the domain important.

After fixing canonicals, structured data, internal links, sitemap scope, and mobile rendering, the next bottleneck is external evidence: useful references from real profiles, publications, datasets, product listings, and community answers.

That is why we are publishing the architecture and a small recipe dataset rather than manufacturing low-quality directory links. The external page must deserve to exist even if the backlink carried no ranking credit.

The resulting 30-recipe image-and-text dataset is available under CC0 on GitHub, with one Markdown record and one image per recipe.

The practical lesson is simple: keep useful pages public, let search engines choose among indexable pages, and use the sitemap to express editorial priority. Those are related decisions, but they are not the same switch.

Explore the live implementation at Fancy Fox.

Disclosure: This article is from the team building Fancy Fox. Recipe concepts and imagery on the service are AI-assisted.

About the author

  • Fancy Fox publishes seasonal recipes, practical cooking ideas, and open recipe data at https://fox.dailyhabit.me/.

No comments yet. Be the first to comment.