Schema markup gets recycled advice because the official documentation is exhaustive and most practitioners copy the first example that seems to fit. After shipping structured data across twenty SaaS sites over the last year, the pattern of what moves CTR and what does not became obvious.

This post is the short version: the six schemas that paid off, the four we'd now skip, and what "working" actually looks like in 2026 now that Google's AI Overviews consume schema alongside traditional results.

The six that worked

1. FAQPage — on pricing + feature pages, not blog posts

Counter-intuitive win. Most SEO guides tell you to slap FAQPage on every blog post. In 2024 that worked; Google demoted that pattern in 2024. The place FAQPage still earns rich results in 2026: pricing pages and individual feature pages, where the Q+A pairs match high-intent queries like "does X include Y" or "how does X handle Z."

What we saw: CTR on pricing-page FAQ schema queries went up 28% on average across five clients. The queries were already ranking; the schema converted position-three visibility into a rich-result position that took meaningful share.

What to avoid: generic blog-post FAQs with "What is X?" openers. Google's current stance is that those are low-quality FAQ patterns and they no longer trigger rich results.

2. SoftwareApplication — for the product landing page

Specific to SaaS. A proper SoftwareApplication schema on your homepage with applicationCategory, operatingSystem, aggregateRating (if you have real reviews), and offers (with a price) gets you the price-in-SERP treatment on product-name queries.

The conversion lift was small but real — ~4% CTR lift on branded queries — and it is one of the cheapest schemas to implement correctly.

3. Article, BlogPosting, and NewsArticle — on the detail page, not the index

Article-type schema on blog detail pages is boring and effective. The fields that matter: author (a real Person with a url pointing to the bio page), datePublished, dateModified, and headline that matches the page's actual H1. Get those four right and you solve 80% of what Article schema is supposed to do.

Where people waste time: adding image arrays with twelve variants, filling in wordCount, adding speakable. None of these move the needle unless you are specifically chasing a voice-assistant use case.

4. BreadcrumbList — non-negotiable for large sites

BreadcrumbList schema is invisible to users but fundamental for crawlers. Every site we audit that had large sections without BreadcrumbList showed signs of crawl inefficiency — pages Google "found" but never indexed. Adding clean BreadcrumbList schema typically recovers those indexing holes within a quarter.

5. Organization — once, on the homepage, with the full profile

Google uses Organization schema to build the knowledge-panel that appears on branded searches. Include logo, sameAs pointing to your real social profiles, contactPoint with a support email, and foundingDate if you have one. This is a set-and-forget.

6. Product — for pricing tables that list discrete tiers

If your pricing page lists three or four discrete plans, each can be a Product with its own offers. This is aggressive but Google accepts it for SaaS in 2026. The win: each plan becomes its own potential rich-result candidate for queries like "Growth tier X" or "pricing comparison X."

One client — an analytics SaaS — saw their Growth-tier pricing page start ranking for "affordable product analytics under $50" after we added per-plan Product schema. That query never ranked before the schema change.

The four we'd skip

Review / aggregateRating without real reviews

Adding aggregateRating with numbers you made up is a manual-action risk. Google's system catches synthetic reviews and demotes the domain. Do not do this. If you do not have review data, omit the field entirely.

HowTo on procedural pages that are actually product flows

HowTo schema was hot in 2023. In 2024 Google shrank its rich-result surface to recipes and a handful of other tight verticals. For SaaS procedural content ("how to set up X integration"), HowTo schema rarely triggers rich results anymore and is a maintenance cost.

VideoObject without a real transcript

Adding VideoObject schema just because you embedded a YouTube video produces no measurable result. The field that matters — transcript — almost no one fills in, and without it the schema is wasted. If you are not going to transcribe, skip it.

LocalBusiness for SaaS

You do not have a local business. Applying LocalBusiness schema to a SaaS site is a miscategorization that sometimes triggers a manual review. If you have a physical office, put it in Organization's address field instead.

How to validate without the drama

Use Google's Rich Results Test — still the fastest sanity check. Validate every deploy of a new schema against it. If the test shows warnings, fix them. If the test shows the rich result preview, you have at least a chance of earning it; Google's final decision depends on ranking and user-intent match.

Schema.org's own validator catches structural errors that Google's test sometimes waves through. Use both, in order: Schema.org first to confirm valid syntax, Google's Rich Results Test second to confirm eligibility.

Implementation pattern

Across our clients the cleanest pattern is:

A shared lib/schema/ module that exports typed builders for each schema.
Every page that needs schema imports the builder and emits <script type="application/ld+json">.
The builders always take real data (no hardcoded fake reviews, no placeholder authors).
A CI check that parses every JSON-LD block on the built site and validates it against Schema.org.

This is all boring. Boring is the point. The schema implementations that quietly work for years share the same boring property: they return real data, in the right shape, on the pages where Google expects to find it.

Where MITPO handles this

The MITPO marketing site ships most of these schemas automatically — Article on blog posts, BreadcrumbList everywhere, Organization on the homepage, FAQPage on the FAQ docs page. If you are using MITPO's hosted blog and landing pages you get those by default; if you are using it for planning and authoring and publishing elsewhere, this post is the checklist for your team to implement on your own platform.

Next step

Turn this into a practical workflow with the competitor research guide.