Featured Snippets & SERP Features

The featured snippet — Google’s “position 0” — was the SERP feature that defined organic strategy for years. AI Overviews has dramatically reshaped its role: many snippets now feed AI Overviews directly, and capturing them gives you visibility in both surfaces. Plus People Also Ask, FAQ rich results, and HowTo schema all share the same structural DNA. Master one pattern and you unlock five SERP features.

TL;DR

• Featured snippets are still alive in 2026, but their role has shifted. Google’s AI Overviews now sit above the snippet on many queries, and snippet content frequently feeds the Overview synthesis. Capture is still worth it — you get cited in both surfaces.
• The four snippet types — paragraph, list, table, video — each have a specific formatting recipe. Match the recipe to the query intent and you have a 30–50% chance of capture against ranked content in positions 4–10.
• PAA, FAQ rich results, and HowTo schema all share the question-and-direct-answer pattern. A page structured well for snippets is structured well for all four features, plus AI Overviews citations and Perplexity source displays.

The mental model

Capturing a featured snippet is like writing a Wikipedia lead paragraph: your goal is to answer the question definitively in the first 40–80 words, in plain language, in a format that can be lifted out and shown standalone. Wikipedia’s lead style is the textbook for snippet-shaped writing.

Each snippet type has a specific shape:

• Paragraph snippet — a 40–80 word direct answer, ideally starting with the query restated as a definition.
• List snippet — an ordered or unordered list with 4–8 items, each item being a short phrase or sentence.
• Table snippet — a true HTML <table> with 2–5 columns and 3–10 rows, headers in the first row.
• Video snippet — a YouTube video with timestamp markers (chapters) that map to the query’s specific moment.

The pattern that wins all four: header phrased as the question, direct answer immediately under it in the matching format, further detail below.

Deep dive: the 2026 reality

The AI Overviews collision. Since the May 2024 launch of AI Overviews and its expansion through 2025, many former featured snippet queries now show an AI Overview instead. The 2025 Ahrefs analysis of 1.4M queries found AI Overviews displaced featured snippets on ~52% of informational queries where both used to appear. But — the AI Overview frequently cites the same source that would have won the snippet, often with the same passage. Capture work pays off in both surfaces.

Snippet types and prevalence (2026 estimates).

Snippet type	Share of snippets shown	Best query types
Paragraph	~70%	“what is”, “why does”, definitions
List (ordered)	~12%	“how to”, step-by-step
List (unordered)	~10%	“best”, “types of”
Table	~6%	comparisons, specs, prices
Video	~2%	“how to”, visual processes

Paragraph snippet recipe. Place a 40–80 word answer in plain HTML <p> immediately after the relevant H2. Restate the query terms in the first sentence. Avoid bullet points or formatting that breaks the lift-out.

List snippet recipe. Use proper <ol> for sequential steps, <ul> for unordered enumerations. Each item should be 5–20 words. Between 4 and 8 items is the sweet spot — Google often truncates after 8.

Table snippet recipe. A real <table> with <thead> and <tbody>. Headers in <th>. Don’t use CSS-styled divs to mimic tables — Google’s snippet extractor pulls from semantic HTML.

People Also Ask (PAA). The PAA accordion has expanded substantially in 2026, often showing 8–12 questions in a self-loading list. Each question Google asks here is a candidate H2 on your page. Capturing one PAA slot for a relevant query both displays your domain in the SERP and is one of the strongest signals to Google that your site is topically aligned with the query.

FAQ rich results. In 2024, Google reduced FAQ rich result eligibility to a much narrower set of authoritative health and government sites — most general sites lost the visible FAQ accordion. FAQPage schema is still useful: it feeds AI Overviews and PAA, and Bing still displays FAQ rich results. The schema is cheap to add; the benefit is asymmetric.

HowTo schema suffered a similar reduction in 2023–2024. Visible HowTo rich results are now rare on desktop and mobile. The schema is still indexed and used by AI surfaces; keep it for genuine procedural content but don’t expect the visible step accordion outside of authoritative sites.

AI Overviews and snippet-shaped content. AI Overviews extracts and synthesizes from multiple sources. Sites with snippet-shaped formatting (question H2, direct answer, supporting detail) get cited more often in the synthesis. Perplexity and ChatGPT Search behave similarly — they extract direct-answer paragraphs and display them as source quotes.

Visualizing it

flowchart TD
  Query["User query: How long should an email subject line be?"]
  Query --> Top["Featured snippet candidates"]
  Top --> P{"Format match?"}
  P -->|"Paragraph in 40-80 words"| Win["Capture paragraph snippet"]
  P -->|"Step-by-step list"| List["Capture list snippet"]
  P -->|"Comparison data"| Table["Capture table snippet"]
  Win --> AIO["Cited in AI Overviews"]
  List --> AIO
  Table --> AIO
  AIO --> Display["Visible in SERP and AI surfaces"]
  Win --> PAA["Question becomes a PAA"]
  PAA --> Display

Bad vs. expert

The bad approach

A wall of text under a generic H2, no direct-answer paragraph, no semantic table.

<h2>About email subject lines</h2>
<p>Email marketers have debated subject line length for years. Some say
shorter is better, others say longer is better. There are many factors
to consider when choosing a subject line length. In this article, we'll
explore everything there is to know about subject line length and help
you find what works for your audience...</p>

<!-- Comparison "table" as nested divs -->
<div class="grid">
  <div class="row">
    <div class="cell">Industry</div>
    <div class="cell">Length</div>
  </div>
  <div class="row">
    <div class="cell">Retail</div>
    <div class="cell">35-40</div>
  </div>
</div>

The H2 doesn’t match a query. The paragraph buries the answer in 90 words of preamble. The “table” is divs, which the snippet extractor ignores. No snippet, no PAA, no AI Overviews citation.

The expert approach

Question-shaped H2, 50–80 word direct answer immediately below, semantic table for comparison data, FAQPage schema.

<h2>How long should an email subject line be?</h2>
<p>Most analyses agree on 41 characters or fewer for English subject lines.
Gmail mobile clips subject lines at 33–43 characters depending on device,
and Litmus's 2025 study of 10,084 emails found subjects under 41 characters
had 22% higher open rates than longer ones. Aim for 30–40 characters for
mobile-first audiences.</p>

<h2>Subject line length by industry</h2>
<table>
  <thead>
    <tr>
      <th>Industry</th>
      <th>Median length</th>
      <th>Top quartile</th>
    </tr>
  </thead>
  <tbody>
    <tr><td>Retail</td><td>38</td><td>32</td></tr>
    <tr><td>SaaS</td><td>45</td><td>38</td></tr>
    <tr><td>Media</td><td>52</td><td>41</td></tr>
    <tr><td>Nonprofit</td><td>48</td><td>40</td></tr>
  </tbody>
</table>

<h2>How to A/B test subject line length</h2>
<ol>
  <li>Pick one product email with consistent send volume above 5,000.</li>
  <li>Create variant A at 30–35 characters and variant B at 50–55 characters.</li>
  <li>Split your list 50/50 randomly using your ESP's native A/B tool.</li>
  <li>Wait 24 hours for the open rate to stabilize.</li>
  <li>Confirm statistical significance with a chi-square test (p &lt; 0.05).</li>
  <li>Apply the winner to your next campaign of similar audience.</li>
</ol>

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "How long should an email subject line be?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Most analyses agree on 41 characters or fewer for English subject lines. Gmail mobile clips subject lines at 33-43 characters depending on device, and Litmus's 2025 study of 10,084 emails found subjects under 41 characters had 22% higher open rates than longer ones."
      }
    },
    {
      "@type": "Question",
      "name": "Does subject line length affect open rate?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes. The Litmus 2025 study showed top-quartile open rates correlated with shorter subject lines, with subjects above 60 characters performing 28% worse than subjects under 40 characters in identical campaigns."
      }
    },
    {
      "@type": "Question",
      "name": "How long should mobile subject lines be?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Aim for 30-40 characters. Gmail's mobile preview cuts at 33-43 depending on device width, and 65% of email opens happen on mobile (Litmus 2025), so optimizing for the mobile preview window matters more than for desktop."
      }
    }
  ]
}

{
  "@context": "https://schema.org",
  "@type": "HowTo",
  "name": "How to A/B test email subject line length",
  "totalTime": "PT72H",
  "step": [
    { "@type": "HowToStep", "name": "Select a recurring email", "text": "Pick one product email with consistent send volume above 5,000." },
    { "@type": "HowToStep", "name": "Build two variants", "text": "Create variant A at 30-35 characters and variant B at 50-55 characters." },
    { "@type": "HowToStep", "name": "Split the list", "text": "Split your list 50/50 randomly using your ESP's native A/B tool." },
    { "@type": "HowToStep", "name": "Wait", "text": "Wait 24 hours for the open rate to stabilize." },
    { "@type": "HowToStep", "name": "Test significance", "text": "Confirm statistical significance with a chi-square test (p < 0.05)." },
    { "@type": "HowToStep", "name": "Apply the winner", "text": "Apply the winning length to your next campaign of similar audience." }
  ]
}

Element	Bad	Expert
H2 phrasing	Generic (“About email”)	Question matching query
Direct answer	Buried after 90 words preamble	First paragraph, 50–80 words
Tables	Divs styled as table	Real <table> with <thead> and <tbody>
Lists	Paragraph mentions	<ol> or <ul> with 4–8 items
Schema	None	FAQPage + HowTo + Article
Capture surface	None	Snippet + PAA + AI Overviews

Do this today

1. In Ahrefs Site Explorer > Organic keywords, filter to SERP features > Featured snippet and People also ask. Sort by Volume. The keywords your site already ranks 4–10 for and that have a featured snippet are your top capture targets.
2. Use Semrush > Position Tracking > Featured Snippets report (or Ahrefs equivalent) to find the queries where a competitor currently owns the snippet but you rank 2–10. These are the steal opportunities.
3. For each target query, identify the snippet type (paragraph, list, table, video) by running the query in incognito Chrome and screenshotting the snippet. Match your formatting to the existing winner.
4. Restructure your page so the target query is an H2 phrased as a question. Place a 50–80 word direct answer in plain <p> immediately below.
5. For list-type targets, convert relevant content to <ol> or <ul> with 4–8 items. Each item: 5–20 words.
6. For table-type targets, convert to a real <table> with <thead> and <tbody>. CSS-styled divs do not capture.
7. Add FAQPage schema for any page with 3+ question-shaped H2s. Validate at validator.schema.org and search.google.com/test/rich-results. The visible accordion is rare in 2026 but the data still feeds AI Overviews.
8. For procedural content (how-to guides, tutorials), add HowTo schema. Same caveat — visible rich result is rare, but AI surfaces and Bing use it.
9. Submit each updated URL via GSC URL Inspection > Request Indexing. Re-check Performance > Search appearance filter > Featured snippet after 14 days to confirm capture.
10. For pages targeting AI Overviews citations, ensure the article also has named author + Person schema + datePublished/dateModified + published expert credentials. AI Overviews preferentially cites pages with strong E-E-A-T markers; the snippet recipe alone is not enough on YMYL queries.