TL;DR
- Markup is a downstream decision. Which type a page deserves is a function of what the page is actually for. Write the answer first. Mark it up second.
- FAQPage markup pairs with a real FAQ block at the bottom of the page. HowTo markup pairs with step-by-step prose. Article markup with a named author bio pairs with E-E-A-T-anchored content. Product or Service markup pairs with entity-disambiguation copy.
- Retrofitting markup onto a page whose content does not match the markup semantic contract produces validation errors and zero citation uplift. The engines do not reward the label. They reward the content the label describes.
- The 1.8x and 1.4x multiples cited for FAQPage markup are vendor-affiliated. The directional signal that markup-matched content lifts more reliably is durable. The exact numbers are orientation, not target.
- The workflow change is the editor’s move. Audit the page’s content shape first. Choose the markup that matches it. Reject any markup the content cannot honestly support.
An in-house team is told that FAQPage markup will lift citation rates. The team takes the announcement at face value. They retrofit FAQPage markup across forty product pages over a sprint.
The validation tools pass. The pages render fine. The marketing director schedules a citation-rate review for the next quarter.
The review comes back flat. Citations did not move. A few pages even dropped slightly. The engines now treat those pages as candidate FAQ surfaces and skip them when looking for definition blocks instead.
The retrofit was correct in execution. The retrofit was wrong in premise.
The pages were not FAQ pages. The team had labelled them as FAQ pages. The label and the content disagreed.
The engines noticed.
Why do markup retrofits rarely lift citation on their own?
Markup is a downstream decision. The engine that lifts a paragraph from your page is reading the paragraph itself. The markup tells the engine what the paragraph is supposed to be. The engine then checks whether the paragraph matches the description.
If the paragraph and the markup agree, the engine ranks the paragraph as a clean candidate for citation. If they disagree, the engine treats the markup as page furniture. It falls back to whatever paragraph in the body looks closest to the question being asked.
The implication for an in-house team is uncomfortable. A markup-first programme retrofits FAQPage, HowTo, or Product types onto pages whose content was never structured for them. The result is validation passes and zero citation uplift. The team ships a quarter of work and watches the metric stay flat.
The 2026 named-source data corroborates the directional finding. GenOptima’s 2026 GEO techniques review reports a 1.8x citation lift in Microsoft Copilot for FAQPage-marked pages and a 1.4x lift in Perplexity. Both multiples are vendor-affiliated.
Frase’s 2026 AEO guide calls FAQPage one of the highest-impact AEO moves. AirOps’s 2026 guide names markup-matched content as a citation input.
The honest hedge. The multiples are orientation, not target. The directional claim — that markup-matched content cites more reliably than markup-mismatched content — is durable across the named sources. The specific numbers will compress as adoption rises industry-wide.
What does "schema follows copy shape" actually mean for the editor?
The editor’s move reverses the order most teams default to. The default order picks the markup type first, retrofits the content into the markup’s contract, and ships.
The reversed order writes the content the buyer needs first. It audits the shape the content has. Then it chooses the markup type whose contract that shape honestly satisfies.
The reversed order produces three benefits. The content reads as natural prose because no markup was forcing it into a shape. The markup validates because the content fits the contract. The engines reward the page because the markup and the content agree.
The reversed order also exposes pages that should not have markup at all. A landing page that is a single sales argument with one CTA does not benefit from FAQPage markup retrofitted at the bottom. The page is not an FAQ.
It does benefit from Article markup with a real author bio. The page has voice and a named writer behind it.
The discipline carries over to revision cycles. When the team rewrites a page, the rewrite changes the content shape. The markup changes after the rewrite ships, not before. The label tracks what is inside the package.
The exception is rare. A team that runs a true Q-A help centre with hundreds of question-and-answer pairs already has FAQ-shaped content. FAQPage markup is the natural label there.
The markup-first instinct happens to land correctly. The exception confirms the rule. The rule is that the content drives the choice.
Which content shape earns FAQPage markup?
FAQPage markup pairs with a real FAQ block. The block carries multiple question-and-answer pairs. Each question is one a real buyer asks.
Each answer fits inside a 40-60 word block. The forty-to-sixty word answer block pillar covers why that length is the load-bearing unit for AEO extraction.
The block sits where buyers expect to find it. Most often that is the bottom of the page. The main argument has been made by then, and long-tail objections and clarifications collect in one place.
The questions are not invented to fit the markup contract. They are the questions support gets. They are the questions sales gets. They are the questions the buyer asks the team after reading the page.
A team that ships an FAQ block built from real buyer questions earns the markup honestly. A team that invents questions to fill the markup gets the validation pass and nothing else.
The pages that deserve FAQPage markup share a pattern. The page exists to answer a question. Or the page exists to make an argument and ends with the questions the argument did not answer in line. Either pattern produces real FAQ content the markup can honestly describe.
The pages that do not deserve FAQPage markup also share a pattern. The page is a sales page. The page is a homepage. The page is a category index.
None of these is FAQ-shaped. The retrofit fails on each because the content was not the kind of content the markup describes.
Which content shape earns HowTo markup?
HowTo markup pairs with step-by-step instructional content. The page walks the buyer through a process from start to finish. Each step is numbered. Each step has a clear input, a clear action, and a clear output.
The reward for matching markup to step-by-step content is double. The engine can present the steps as a sequence in answer surfaces. The buyer who lands on the page sees the same sequence visually. The two audiences get the same shape from the same prose.
The honest test for HowTo eligibility. Read the page top to bottom. Is the content a sequence of steps that produces a result if followed in order?
If yes, HowTo markup fits. If no, HowTo markup does not fit no matter how many headings the page has.
A common retrofit failure. A team marks a list of features with HowTo because the list has bullets. The list is not a sequence.
The bullets are not steps. The buyer does not produce a result by reading the list in order. The markup fails because the content is not what the markup describes.
The fix is to write the steps. Or to drop the markup. Either is honest. Both ship.
Why does Article markup with a real author bio carry E-E-A-T weight?
Article markup pairs with content that has a writer behind it. The writer has a name. The writer has credentials. The writer has a bio that the page renders for any reader who wants to know who wrote what they are reading.
GenOptima’s 2026 review cites a Search Engine Land study. Pages with named authors and full bios were cited 2.3x more often than pages without. The number is a secondary citation.
The original study is the primary if the claim is load-bearing. The directional signal stands — named authors with bios outperform anonymous content. AirOps corroborates the finding, noting "visible bios, credentials, and experience signals" as inputs to citation likelihood.
The implication is that the author bio is a copy job, not a markup job. The Article markup describes the author. The bio is the content the markup describes.
A bio that names credentials, experience, and a real point of contact carries the weight. A bio that says "the team" or "editors" does not.
The discipline that follows. Every editorial-grade page on the site has a named author. Every named author has a bio page on the site. Every bio page links the author to verifiable signals — published work elsewhere, organisational memberships, real credentials.
The Article markup describes that infrastructure honestly. The engine rewards the page because the description is real.
A team that wants the citation lift but cannot stand up the author infrastructure should not ship the markup. Marking up content with an author who has no bio invites the engine to discount the page. The markup contract was not honoured.
What happens when markup is forced onto the wrong content shape?
Forcing markup onto a mismatched page produces three failure modes. Validation tools may pass — the engines reading the page do not.
The first failure mode is validation errors that escape the team’s tooling. Most teams audit markup with a single validator. The validator passes pages that meet the strict semantic contract. It misses pages whose content drifts from the contract over time.
A page that started as an FAQ block and was edited into a sales pitch retains the FAQPage markup. The validator passes. The engine penalises the mismatch.
The second failure mode is silent citation suppression. The engine reads the markup, expects a particular content shape, and finds something different. It skips the page when looking for that shape elsewhere on the web.
The team sees no error. The team sees no improvement. The work disappears into a non-result.
The third failure mode is the one most teams notice last. Forced markup damages the surrounding content because the team rewrote real prose to fit the contract. A landing page that became three FAQ entries to satisfy FAQPage markup lost the persuasive arc that closed sales for it. The retrofit cost the team the conversion the page existed to drive.
The honest fix is reversal. Audit which pages have markup. Audit whether the content matches. Where the content does not match, choose: rewrite the content to match the markup, or strip the markup.
Both are valid. Shipping the mismatch is not.
How does the workflow change when the copy is written first?
The workflow change is small in process and large in outcome. The team writes the page first. The team reads the page top to bottom.
The team identifies what shape the page actually is. The team chooses the markup type whose contract that shape satisfies. The team ships markup and content together.
The change cascades through three places. Editorial calendars now name the content shape before naming the markup type. Markup decisions move from the production checklist to the upstream planning document. Validation runs against content fidelity, not just contract compliance.
The benefit shows up in two metrics. Citation rates rise on pages where markup-content agreement was historically poor. Validation errors drop because the contracts the pages now claim are contracts the pages can honour.
The cost is a slower production cadence in the first quarter. Teams used to retrofitting markup to existing pages now spend audit time reversing the order on the back catalogue. The cost amortises across the next quarter as the discipline becomes default.
A senior editor can lead the discipline shift without restructuring the team. The shift is a checklist change, not a roles change. The checklist now reads: identify shape, choose markup, ship together. The team that runs the checklist consistently earns the lift the markup retrofits never delivered.
Other questions worth answering
How do voice assistants treat Speakable annotations laid over plain-prose answers?
Voice surfaces favor Speakable annotations laid over short, plain-prose answers. The 2026 Copywriting for AEO surfaces research narrative names Speakable as a niche but real lever for brands whose buyers ask questions out loud. The reward shows up only when the spoken-shaped sentence already exists in the prose.
Does community-sourced ranking inside Perplexity reward different prose than encyclopedic ranking inside ChatGPT?
Yes. Per the 2026 Copywriting for AEO surfaces research, Reddit supplies roughly 6.6% of Perplexity sources and 46.7% of its top-10. Wikipedia supplies 7.8% of ChatGPT sources and 47.9% of its top-10. One rewards peer voice, and the other rewards definition-led factual prose.
Where does Microsoft Copilot reward table-shaped prose more than paragraph-shaped prose?
Copilot rewards tables when the buyer is comparing two or more options across the same attributes. The 2026 Copywriting for AEO surfaces research notes Copilot cites 28% of answers and leans on Bing-indexed table data. A paragraph that buries the comparison loses to a row-and-column version of the same facts.
When does an unnamed byline cost a brand measurable visibility in the engines?
An unnamed byline costs visibility whenever the buyer-facing question touches expertise, judgment, or trust. The 2026 Copywriting for AEO surfaces research cites a Search Engine Land finding of a 2.3x lift for named bylines with full bios. The vendor-bias caveat applies. The directional signal is durable.
Where would you start auditing the next page on your site?
Pick a page that has markup today and is not citing in the engines you actually care about. Open the page. Read it top to bottom. Ask one question: what shape does the content actually take?
If the content is a single argument with a CTA, the page is a sales page. Article markup with a real author bio is the honest match. FAQPage and HowTo are mismatches. If either is on the page today, strip them.
If the content is a list of steps that produces a result when followed, the page is a how-to. HowTo markup is the honest match. Article markup is fine in addition. FAQPage is a mismatch.
If the content is a real FAQ block built from buyer questions, FAQPage markup is the honest match. The block belongs at the bottom of the page when the page also has a primary argument. The block belongs as the page itself when the page is a help-centre entry.
If the content does not match any of the markup types, ship the page without markup beyond Article. Article with a named author bio is the floor for any editorial-grade page on a site that wants to cite well. The other markup types layer on top when the content earns them.
The framework you reach for to structure the prose underneath the answer block is a separate decision. Some copywriting frameworks for beginners compress cleanly into the 40-60 word block the engine lifts. Others lose load-bearing structure under that compression.
The framework choice is independent from the markup choice. Both decisions follow the content shape, not the other way around.
If you would like another set of eyes on the markup-content audit before you ship the rewrite, you can reach me here. I will read the pages against the engines I track and the markup contracts they claim. The first round is free, and the audit is yours regardless of whether we work together after.