Existing markup audit
What structured data you currently emit, whether it validates, and whether it still matches the visible content. Mismatches are common after any content change.
What this markup does now that rich results are gone
It defines answer boundaries. Google largely stopped showing FAQ rich results for most sites in 2023, so the visible reward disappeared. What remains is machine-readable structure: an engine parsing your page knows precisely which text answers which question, rather than inferring it.
This is where most published advice is wrong, in both directions. One camp still promises rich snippets that no longer appear for most sites. The other concluded the markup is worthless and stripped it out. Both misread what changed — the display was withdrawn, not the parsing.
The parsing matters more now than it did when rich results existed, because answers are increasingly synthesised rather than displayed verbatim. A system assembling an answer from several sources benefits from knowing unambiguously where your answer to a specific question starts and stops. Markup provides that; inference is less reliable.
HowTo markup has followed the same path and retains the same residual value. The step structure it declares is genuinely useful to any system trying to extract a procedure, whether or not a visual carousel is ever shown for it.
The failure mode we see most often is markup that does not match the page. Teams add FAQPage entries for questions the page never really answers, or leave markup behind after the content changed. Invalid or mismatched structured data is worse than none, because it teaches an engine that your declarations are unreliable.
What the implementation covers
The implementation covers an audit of existing markup and its validity, FAQPage and HowTo schema on the pages where it genuinely fits, alignment between markup and visible content, validation across your templates, and a governance rule so it does not drift out of sync again.
FAQPage implementation
Applied to pages that genuinely answer distinct questions, with each answer marked to its actual boundary rather than to a convenient block of text.
HowTo implementation
Step structure declared for genuine procedures, with the sequence and any tools or prerequisites captured properly rather than flattened into prose.
Content alignment
Markup reconciled with what the page visibly says, because declaring an answer that is not present is the fastest way to have your structured data discounted.
Template-level validation
Validation wired into the templates that generate your pages, so new content inherits correct markup instead of relying on someone remembering.
Drift governance
A rule and a check for keeping markup in sync when content changes, since this is the point at which most implementations quietly become inaccurate.
How the implementation runs
The implementation runs over three to five weeks: audit what you currently emit and whether it validates, decide which pages genuinely warrant markup, implement at template level rather than page by page, then validate and put a drift check in place.
- 01
Audit what exists
Current structured data across templates, its validity, and whether it still matches visible content. Legacy markup left behind by past content changes is extremely common.
→ A validity and mismatch report
- 02
Decide where it belongs
Pages that genuinely answer distinct questions or describe real procedures. Adding FAQPage markup to a page with no real questions is a common way to make things worse.
→ An agreed markup scope by template
- 03
Implement at template level
Generated from the content itself wherever possible, so new pages inherit correct markup automatically instead of depending on an editor remembering to add it.
→ Template-level markup, deployed
- 04
Validate and govern drift
Validation across templates plus a check that flags when content changes leave markup describing something the page no longer says.
→ Validated markup and a drift check
What changed in 2023 and what did not
Google withdrew FAQ and HowTo rich results for most sites, so the visible reward for this markup largely disappeared. The machine-readable value did not change at all, and arguably increased as answers moved from being displayed verbatim to being synthesised from several sources.
| Before 2023 | Now | |
|---|---|---|
| Rich result display | FAQ accordions shown in results | Withdrawn for most sites |
| Why teams implemented it | Visible SERP real estate | Reliable answer extraction |
| Answer boundary signal | Useful | Useful, and more so as answers are synthesised |
| Cost of invalid markup | Rich result withheld | Your declarations get discounted generally |
| Common mistake | Marking up everything for the display | Stripping it out because the display went |
Signals you need this now
You need this when your structured data fails validation, when markup describes answers the page no longer contains, when FAQ markup was stripped out after the rich results went, or when new pages ship with no markup because nobody remembered to add it.
- Structured data reports validation errors in Search Console
- Markup describes answers the page no longer contains
- FAQ markup was removed entirely after rich results were restricted
- New pages ship without markup because it is added manually
- Your FAQ blocks have no schema at all
- How-to content declares no step structure
- Nobody can say which templates emit which markup
What to expect, realistically
Schema is a supporting signal rather than a lever that moves rankings by itself. Expect cleaner extraction, fewer answers truncated mid-sentence, and validation errors resolved. Anyone promising ranking gains from markup alone is selling something the evidence does not support.
Where this fits
Related AEO services
Common in these sectors
The adjacent disciplines
Deciding between options
Questions about FAQ schema
Each answer is written to stand alone in 40 to 60 words — the shape an AI Overview or Perplexity citation lifts. Ships with FAQPage schema.
Is FAQ schema still worth implementing after 2023?
Yes, for a different reason than before. Google withdrew the rich result display for most sites, but the markup still tells engines exactly where each answer starts and ends. That matters more now that answers are synthesised from multiple sources rather than displayed verbatim.
Will FAQ schema improve our rankings?
Not directly. It is a supporting signal that makes extraction cleaner and reduces answers being truncated. Anyone promising ranking gains from structured data alone is overstating what the evidence supports, and we would rather say so than sell it that way.
Should we add FAQ markup to every page?
No. It belongs on pages that genuinely answer distinct questions. Adding it to pages without real questions produces mismatched markup, which is worse than none — it teaches engines that your structured declarations are unreliable in general.
What happens if our markup does not match the page?
It gets discounted, and the discounting is not limited to that page. Declaring answers the content does not contain undermines trust in your structured data generally, which is why we implement at template level and add a drift check.
Is HowTo markup still useful?
Yes, for the same reason as FAQPage. The visual carousel was withdrawn, but the declared step structure remains genuinely useful to any system extracting a procedure. Implement it where you have real procedures, not to decorate prose.
Can our developers implement this themselves?
Usually, and we are happy to specify rather than build. The part worth paying attention to is generating markup from the content itself at template level, so it cannot drift out of sync the next time someone edits a page.
Find out whether this is your constraint.
Thirty minutes with a senior strategist. We pull your live visibility while we talk and tell you plainly whether a FAQ schema is what you need — or whether your problem sits somewhere else.