Hard-coded JSON-LD drifts out of date. Generate it from the same data as your pages, and keep only the schema types Google still uses.
Introduction
Explaining your SaaS to a human is easy: a screenshot, a 30-second demo, some punchy copy. Done.
A crawler gets none of that. It reads text and markup. Structured data, usually written as JSON-LD (JavaScript Object Notation for Linked Data), is the part of your page that states in plain terms what the page is: who published it, what it sells, when it was last updated.
Be realistic about what it buys you. Google says its AI features need no [special schema.org markup](https://developers.google.com/search/docs/appearance/ai-features), so schema is not a requirement for being cited. It is context a machine can read without guessing, and Google [recommends JSON-LD](https://developers.google.com/search/docs/appearance/structured-data/intro-structured-data) as the easiest format to maintain. The one thing that matters most is that it stays accurate.
The "Stale Data" Trap of React Components
I see this one in almost every vibecoded project I look at.
A developer reads about SEO, opens `index.html` or their root `App.tsx`, and drops in a static script tag like this:
```html <script type="application/ld+json"> { "@context": "https://schema.org", "@type": "SoftwareApplication", "name": "My App", "offers": { "@type": "Offer", "price": "10" } } </script> ```
They ship it and move on.
Two weeks later marketing bumps the price to $19. Or they rename the product. The visible page updates, but that hidden JSON blob still says $10.
Now your markup and your page disagree. Google's [general structured data guidelines](https://developers.google.com/search/docs/appearance/structured-data/sd-policies) say markup must describe content that is visible on the page. A stale price breaks that, and any tool that reads both has a reason to trust neither.
The rule: never hard-code metadata. Generate it at build time from the same source the rest of your app reads from.
Which schema types are worth keeping accurate
Schema.org has hundreds of types. Most of them don't matter for a SaaS site. Three groups are worth the effort.
**Organization and WebSite.** These tell search engines who you are: your name, logo, URL and the pages that represent you. They help a search engine identify your brand. Put them on your homepage and keep them accurate.
**Article, BlogPosting or TechArticle.** For posts and docs, mark up the headline, the author and the published and modified dates. The dates must match the dates a reader sees on the page. `TechArticle` adds fields such as `proficiencyLevel`, which is fine to include if it is true, but no search engine we know of documents using it to choose what to show a beginner.
**SoftwareApplication.** If you ship a tool, this describes what it is, its category and its price. Keep the price in step with your pricing page, which is the stale-data trap above.
**Skip the retired types.** FAQPage no longer earns a Google rich result: that ended on 7 May 2026. HowTo is retired. The sitelinks search box, which `SearchAction` powered, was removed in 2024. You can keep FAQ markup if it matches questions that readers can see on the page, but don't build a strategy on it. You can follow these changes on Google's [Search documentation updates](https://developers.google.com/search/updates) page.
One more correction while we are here. Schema.org has no type for a table of contents. If a tool promises to generate one, it is inventing vocabulary that no parser reads.
Why you want this automated
Doing this by hand on every page isn't realistic. It is tedious, it breaks the moment marketing edits a price, and nobody remembers to update it.
Treat schema as part of your build. Keep one data file for the things that appear in more than one place: your name, URL, logo, prices, dates and authors. Render the page from it, and render the JSON-LD from it too:
```ts const jsonLd = JSON.stringify({ "@context": "https://schema.org", "@type": "Organization", name: site.name, url: site.url, logo: site.logo, }); // rendered into the page head at build time const tag = `<script type="application/ld+json">${jsonLd}</script>`; ```
Now the structured version of your site changes whenever the visible version changes. There is no drift and no $10-versus-$19 problem.
One last thing
JSON-LD is how you tell a machine what a page is, in a format it can parse. Keep it small, keep it true and generate it from the same data as the page.
Want to know which types your homepage publishes and whether the JSON-LD parses? Our free visibility report reads every JSON-LD block on your homepage and lists the types it finds: [scan your site for structured data now.](https://cite-met.com)
Scanner checks this post covers
- CM-902 Structured data: Structured data is JSON-LD markup that describes your page in schema.org terms.