# AIO vs Snippet Decision Framework

The strategic question for 2026: is this a snippet-target keyword, or is it an AIO-target keyword? Chasing the wrong one wastes the rewrite.

## The Current Reality

AI Overviews now appear on roughly 13% of all queries globally ([Ahrefs, 38% of AI Overview Citations Pull From The Top 10](https://ahrefs.com/blog/ai-overview-citations-top-10/)), with higher rates in specific verticals (Science 43.6%, Health 43.0%, People & Society 35.3%). Some studies using different metrics report AIO on 25.8% of US searches or 58% of informational queries specifically. The range reflects how you slice the data — but the direction is the same: AIO is growing and it's displacing featured snippets where the two overlap.

The correlation between featured snippet decline and AIO growth is 0.9 ([Ahrefs, Goodbye Featured Snippets](https://ahrefs.com/blog/how-serp-features-have-evolved-in-the-ai-era/)). Featured snippet prevalence dropped from 15.4% of queries to 5.5% as AIO expanded through 2025. Featured snippets and AIOs rarely co-occur on the same query — when Google shows one, it tends to suppress the other, and that separation has widened as AIO coverage has grown.

What this means in practice: a query that used to reliably show a snippet might now show an AIO, or neither. Before rewriting, always Google the query and see what's actually there.

## Three Decision Criteria

Before committing to snippet optimization, answer these in order:

### 1. Is there currently a featured snippet for this query?

Google the query. If yes, the snippet game is still in play for this keyword. Proceed.

If no, go to criterion 2.

### 2. Is there an AI Overview?

If yes, snippet optimization is secondary. The primary target becomes getting cited inside the AIO, because brands cited in AI Overviews earn 35% more organic clicks than non-cited brands ([Ahrefs / Seer Interactive, December 2025](https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/)).

If neither a snippet nor an AIO is present, go to criterion 3.

### 3. What's the query type?

- **Short definitional or factual queries** (what is X, when did X, who invented X) — snippets still reliably show. Proceed with snippet optimization.
- **Opinion, "best X", and comparison queries** — these have been absorbed by AIO in most verticals. "Best X" queries shifted almost entirely to AIO by mid-2025. Focus on AIO citation, not the snippet.
- **Procedural "how to" with a concrete task** — snippets still hold for specific tasks (how to change a flat tire, how to reset a router). AIO is less aggressive on procedural content.
- **"Why" questions** — reason queries trigger AIO 59.8% of the time (Ahrefs), the highest rate of any category. Treat these as AIO-target keywords by default.

## The Four-Quadrant Matrix

|  | No AIO | AIO Present |
|---|---|---|
| **Snippet present** | Target the snippet. Highest ROI — clean single-target rewrite. | Dual optimization: target both. Write for the snippet (H2 + 40-60 word block), add structured data, ensure the top 10 ranking. |
| **No snippet** | Target the SERP feature Google is likely to add next — usually "People Also Ask" (PAA) or related searches. Add a FAQ section with PAA-style questions. | Target AIO citation, not a snippet. Focus on entity-rich writing, structured data, answer-first format. |

Each quadrant has a different rewrite brief. Confusing them is the main reason snippet optimization projects fail.

## How to Target AIO Citations (Brief)

When the decision lands on "AIO target", the playbook is different from snippet optimization:

- **Structured data** — FAQ schema has a 3.2x higher AIO citation rate. HowTo schema works for procedural content. Both signal to Google that the content is extractable.
- **Answer-first format** — the same 40-60 word answer blocks that win snippets also win AIO citations. 92% of AI Overview citations come from top-10 ranking pages, so the fundamentals of ranking still apply.
- **Entity-rich writing** — name the specific brands, tools, people, and concepts. AIO citation models extract entities and match them to queries; vague prose gets skipped.
- **E-E-A-T signals** — author bylines, cited sources, update dates. AIO citations skew toward pages with clear authorship.

For the full AIO citation playbook, see `skills/write-content/references/geo-optimization.md`.

## Realistic Timelines

- **Snippet changes** — Google reprocesses snippets within 1-4 weeks of a content change, assuming the page is already indexed and ranks in the top 10. If there's no change in 4 weeks, recheck your H2 phrasing and answer length.
- **AIO citations** — less predictable. Citation changes can take 4-8 weeks, and the AIO regenerates on each query without a stable cache. You'll see citations come and go.
- **Don't iterate too fast** — after a rewrite, wait at least two weeks before making further changes. Faster iteration confuses the measurement.

## Cross-Reference: Zero-Click Assessment

Before committing to either snippet optimization or AIO citation targeting, run the zero-click assessment from the keyword-deep-dive skill: `skills/keyword-deep-dive/references/zero-click-and-aio-impact.md`. That file estimates how much traffic the keyword can actually produce given current SERP features. If the zero-click rate on a query is 80%+, even winning the snippet or the AIO citation won't move the clicks needle much — and the effort is better spent elsewhere.

## A Short Checklist

Before starting a rewrite, answer:

1. Did I Google the query today? (Not last month. Today.)
2. Is there a featured snippet?
3. Is there an AI Overview?
4. What's the query type? (Definitional / how-to / comparison / opinion / why)
5. What does the zero-click assessment say about traffic potential?
6. Which quadrant am I in?
7. What does the correct quadrant playbook say to write?

If any answer is "I don't know", stop and find out. The cost of optimizing for the wrong SERP feature is writing the wrong content and getting zero snippet wins.

# Query Format Matching — Expanded

Google picks a snippet format based on query shape. Write the answer in the wrong format and another page will win, even if your content is better. This is the full pattern list.

## The Expanded Table

| Query Pattern | Snippet Format Google Shows | Example Query | Notes |
|---|---|---|---|
| "What is X" | Paragraph, 40-60 words | "what is affiliate marketing" | Most common snippet type. Answer sits directly under an H2 that mirrors the query. |
| "What does X mean" | Paragraph, 30-50 words | "what does SEO mean" | Shorter than "what is X" — Google wants a dictionary-style gloss, not an explainer. |
| "What are X" | Paragraph or unordered list | "what are backlinks" | Splits on subject: single definition → paragraph; plural enumerable → list. |
| "How to X" | Ordered list, 3-8 steps | "how to bake sourdough" | The canonical list-snippet query. Steps must use `<ol>` with `<li>`. |
| "How do I X" | Ordered list, 3-8 steps | "how do I reset my router" | First-person variant, same format. |
| "How do you X" | Ordered list, 3-8 steps | "how do you make cold brew" | Second-person variant. Same rules. |
| "Steps to X" | Ordered list, 5-10 steps | "steps to start an LLC" | Longer lists tolerated here — Google expects procedural depth. |
| "X tutorial" | Ordered list | "python tutorial" | Same as how-to; tutorial implies sequence. |
| "X vs Y" | Table (primary), paragraph (fallback) | "react vs vue" | Table wins when the page uses a real `<table>` element with clear column headers. Without a table, Google falls back to a comparison paragraph. |
| "X or Y" | Paragraph, opinionated | "css grid or flexbox" | Different from "vs" — reads as a decision, not a spec sheet. One-paragraph recommendation format. |
| "Difference between X and Y" | Paragraph or table | "difference between ram and rom" | Paragraph is more common; table only if both entities have multiple parallel attributes. |
| "Types of X" | Unordered list, 5-10 items | "types of cheese" | Each item = one noun phrase, sometimes with a 5-10 word gloss. |
| "X list" | Unordered list | "programming languages list" | Same format as types-of, usually larger (8-12 items). |
| "Best X" | Unordered list with brief descriptors | "best running shoes" | The most AIO-displaced pattern — see reality check below. |
| "Top X" | Unordered list | "top email marketing platforms" | Same as "best X". Often numbered in the source but Google still renders as bullets. |
| "When did X" | Short paragraph, under 40 words | "when did rome fall" | Direct date/year answer with 1-2 sentences of context. |
| "When was X" | Short paragraph | "when was the internet invented" | Similar to "when did X"; includes a date and actor. |
| "Who is X" | Short paragraph, person bio format | "who is elon musk" | 2-4 sentences, usually pulled near a headshot or knowledge panel. |
| "Who invented X" | Short paragraph | "who invented the telephone" | Name + year + one-line context. |
| "Why does X" | Paragraph, explanation format | "why does iron rust" | 40-70 words. Causal structure: mechanism + outcome. |
| "Why is X" | Paragraph | "why is the sky blue" | Same as "why does X". "Why" question words have historically had the highest snippet trigger rate of any question word — optimize as if a snippet is available. |
| "How much does X cost" | Table (ranges) or paragraph (single price) | "how much does a tesla cost" | Price ranges across variants → table. Single product → paragraph with the number up front. |
| "How long does X take" | Short paragraph with specific time | "how long does a passport take" | Answer the duration in the first clause; qualify in the second. |
| "How many X" | Short phrase, single number | "how many bones in the human body" | Sometimes just a number card. Format the answer as "There are N X" in the first sentence. |
| "Where is X" | Short paragraph with location | "where is machu picchu" | Often paired with a map card, not a classic snippet. Still write a paragraph answer under an H2. |
| "X definition" | Paragraph, dictionary format | "paradox definition" | Shorter than "what is X" — 20-40 words, dictionary tone. |
| "X examples" | Unordered list or paragraph with inline examples | "metaphor examples" | List format wins when examples are parallel and short. Paragraph wins when each example needs context. |
| "Is X [adjective]" | Paragraph, direct yes/no | "is coffee bad for you" | Start with "Yes" or "No", then 2-3 sentences of qualification. Google rewards the literal yes/no opener. |
| "Can X Y" | Paragraph, yes/no | "can dogs eat chocolate" | Same rule as "Is X" — commit to a direct answer first word. |
| "X calculator" | Calculator widget, not a snippet | "mortgage calculator" | Not a snippet target at all. Google serves an interactive calculator. Don't chase this. |
| "X meaning in [language]" | Paragraph, translation format | "amigo meaning in english" | Often replaced by a translation card — snippet is secondary. |

## Ambiguous and Edge Cases

- **"How much X"** (no "cost", no "take") — Google reads intent on a case-by-case basis. "How much sleep do I need" shows a paragraph. "How much is a Tesla" shows a table. When in doubt, write both: a paragraph block directly under the H2 and a table further down. Google picks.
- **"Best X for Y"** — the sub-qualification ("for beginners", "for small business") sometimes shifts Google to a substructured list with named categories ("Best overall: X. Best budget: Y. Best for beginners: Z."). Target this format when the sub-qualifier adds a dimension that splits the list cleanly.
- **Comparison queries with 3+ options** — "react vs vue vs angular" theoretically wants a table, but Google frequently rejects the snippet entirely and shows an AI Overview instead because the comparison surface is too wide for a clean snippet. When you see an AIO on a 3-way comparison, stop chasing the snippet and optimize for the AIO citation instead.
- **Questions with an implicit list** — "what are the ingredients in coca cola" reads as a what-is query but Google often shows an unordered list. Test the SERP; don't trust the query stem.
- **"X near me"** — local intent, never a classic snippet. Local pack only. Skip.

## 2026 Reality Check

Featured snippet prevalence dropped from 15.4% of queries to 5.5% as AI Overviews expanded ([Ahrefs, Goodbye Featured Snippets](https://ahrefs.com/blog/how-serp-features-have-evolved-in-the-ai-era/)). A large-scale Ahrefs study recorded a 64% drop in featured snippet visibility between January and June 2025 alone, correlating 0.9 with AI Overview growth. The takeaway: many query patterns in the table above no longer trigger a snippet at all.

Which query types still reliably show featured snippets in 2026:

- **Short factual/definitional queries** — "what is X", "when did X", "how many X", "who invented X". These still show snippets because AIO is overkill for a one-line answer.
- **How-to queries with concrete tasks** — "how to change a flat tire", "how to reset X". Procedural content holds up; AIO is less aggressive here.
- **Specific numeric questions** — "how long does X take", "how much does X cost" for a single item. Single-number answers resist AIO replacement.

Which query types have been absorbed by AI Overviews:

- **"Best X" and "top X"** — these shifted almost entirely to AIO by mid-2025 because they benefit from multi-source synthesis.
- **"Why" questions** — reason queries trigger AIO 59.8% of the time (Ahrefs), the highest rate of any category. The snippet you used to see is probably now an AIO.
- **Comparison queries** — "X vs Y" with 3+ options or complex dimensions usually show AIO instead of a table.
- **Long-tail informational queries** — the longer and more synthesis-heavy the query, the more likely AIO replaces the snippet. Rule of thumb: past 6-8 words with multiple clauses, assume AIO.

Rule of thumb: if the query is longer than 6 words or demands synthesis across sources, assume AIO. If the answer fits in a tweet, the snippet is still in play.

## How to Use This Table

1. Look at the target query.
2. Find the closest row in the table.
3. Write the answer in that exact format, under an H2 that mirrors the query.
4. Before committing, Google the query and check if a snippet actually shows. If it's an AIO instead, re-scope — see `aio-vs-snippet-decision.md`.

# Snippet Format Templates — Worked Examples

Each format below has three failing versions paired with the rewrite that would win the snippet. Copy the structure, not the content.

## Paragraph Snippets (40-60 words)

Paragraphs are 70% of all featured snippets ([Semrush/Ghergich study](https://www.semrush.com/blog/featured-snippets/)). The target is 40-60 words, with 45 words being the statistical sweet spot. Every example below is a real failure pattern — answer buried, format wrong, or structure missing the H2-then-answer handoff.

### Example 1: "what is content marketing"

**Failing version** (a 300-word paragraph buried under an H3 on the fourth scroll of a marketing blog):

> ### The rise of content marketing
> In today's digital landscape, marketers are constantly searching for new ways to reach their audiences. One approach that has gained significant traction over the last decade is content marketing. But what exactly does that mean? To understand content marketing, we first need to look at how traditional advertising has failed... [continues for 14 more sentences before the actual definition appears]

**Winning rewrite** (new H2, 47 words, immediately under the heading):

> ## What is content marketing?
> Content marketing is a strategy where businesses create and share articles, videos, and other media to attract and retain customers. Instead of directly promoting a product, content marketing builds trust by providing useful information that answers questions the audience already has.

Word count: 47. Keyword in the first five words. Self-contained definition. The rest of the page can now expand into history, examples, and tactics — but the snippet block owns the top of the section.

### Example 2: "what does SEO stand for"

**Failing version** (the acronym explained three paragraphs into an article, after an anecdote):

> When I started my first blog in 2008, I had no idea what I was doing. I spent weeks writing posts that nobody read. Eventually a friend told me I needed to learn about SEO. At first I thought SEO was a made-up word. It turns out SEO stands for Search Engine Optimization, which is the practice of...

**Winning rewrite** (new H2, 25-word answer block):

> ## What does SEO stand for?
> SEO stands for Search Engine Optimization. It's the practice of improving a website so it ranks higher in search engines like Google for relevant queries.

Word count: 25. The first sentence is the literal answer. The second adds one line of context. No narrative, no preamble.

### Example 3: "what is a backlink"

**Failing version** (the definition mixed into a glossary section listing 40 SEO terms):

> **Backlink** — See also: inbound link, external link. A backlink is when another website links to your website. This is important for SEO, as discussed in the section on link building above. Related: dofollow, nofollow, anchor text.

**Winning rewrite** (standalone H2, 45 words, no cross-references):

> ## What is a backlink?
> A backlink is a link from another website that points to your website. Search engines treat backlinks as signals of authority — pages with more high-quality backlinks tend to rank higher. A backlink is also called an inbound link or external link.

Word count: 45. The definition is self-contained. A reader who lands on the H2 understands the concept without scrolling up.

## Ordered List Snippets (How-To)

Ordered lists work for "how to", "how do I", and "steps to" queries. Google pulls 3-8 steps; anything longer gets truncated. Each step needs to be a complete action the reader can execute without reading the previous step in full.

### Example 1: "how to change a flat tire"

**Failing version** (prose narrative in six paragraphs, no `<ol>`, no action verbs at the start):

> The first thing you want to do is make sure your car is in a safe location. Pull onto a flat, stable surface, away from traffic. Then you'll need to locate your spare tire and jack, which are usually in the trunk...

**Winning rewrite** (H2, six numbered steps with action verbs, one sentence each):

> ## How to change a flat tire
> 1. Park on a flat surface and engage the parking brake.
> 2. Remove the hubcap and loosen the lug nuts with a wrench.
> 3. Place the jack under the car's frame and raise the vehicle.
> 4. Unscrew the lug nuts and remove the flat tire.
> 5. Mount the spare tire and hand-tighten the lug nuts.
> 6. Lower the car, fully tighten the lug nuts, and store the flat.

Six steps, each under 15 words, each starting with an action verb. The `<ol>` structure is what Google extracts.

### Example 2: "how to start a blog"

**Failing version** (15 steps, each with paragraph-length explanations):

> 1. First, you need to find your niche. Think about what you're passionate about, what you know, and what has an audience. Spend at least a week researching before you commit...
> 2. Next, brainstorm a domain name. This is crucial because your domain is your brand...
> [continues through step 15]

Google truncates lists at 8 items and each item at ~20 words. 15 steps with paragraph explanations lose on both counts.

**Winning rewrite** (5 steps condensed, each under 20 words):

> ## How to start a blog
> 1. Pick a niche you know well and that has a searchable audience.
> 2. Buy a domain name and sign up for a hosting provider.
> 3. Install WordPress or another content management system.
> 4. Choose a theme and write your first three articles.
> 5. Publish and share on one social channel where your audience already hangs out.

Five steps, all under the truncation limit. The long version can still live on the page as a deeper explainer below the snippet block.

### Example 3: "how to meal prep"

**Failing version** (steps that run to 40+ words each):

> 1. Start by planning your meals for the week. Sit down on Sunday morning with a notebook and think about what you've eaten recently, what's on sale at the grocery store, what ingredients you already have in your pantry, and what your nutrition goals are for the coming week.
> 2. Next, make a shopping list based on those meals...

**Winning rewrite** (steps under 20 words each):

> ## How to meal prep
> 1. Plan five meals for the week and write down the ingredients.
> 2. Shop for everything in one trip to save time.
> 3. Cook all proteins and grains in one 90-minute session.
> 4. Portion meals into containers and refrigerate or freeze.
> 5. Reheat each day and eat within five days.

## Unordered List Snippets

Unordered lists fit "types of X", "best X", and "signs of X" queries. Same rules as ordered lists — 5-8 items, short phrasing, parallel structure.

### Example 1: "types of coffee drinks"

> ## Types of coffee drinks
> - Espresso
> - Americano
> - Latte
> - Cappuccino
> - Macchiato
> - Mocha
> - Cortado
> - Flat white

Eight items, each one word (or two), all nouns. Google renders this directly.

### Example 2: "best running shoes for beginners"

> ## Best running shoes for beginners
> - Brooks Ghost 16 — cushioned neutral shoe for road running
> - Hoka Clifton 10 — maximum cushioning with a smooth heel-to-toe transition
> - Asics Gel-Kayano 32 — stability shoe for mild overpronation
> - New Balance 880v15 — durable daily trainer for high mileage
> - Nike Pegasus 41 — versatile neutral shoe that handles speed work
> - Saucony Ride 18 — balanced cushioning for new runners

Six items. Brand name first, then a 5-8 word descriptor. The dash separator reads cleanly in Google's snippet renderer.

### Example 3: "signs of burnout"

> ## Signs of burnout
> - Chronic exhaustion that sleep doesn't fix
> - Cynicism or detachment from work
> - Reduced professional effectiveness
> - Frequent headaches or muscle pain
> - Difficulty concentrating on simple tasks
> - Loss of motivation for things you used to enjoy
> - Increased irritability with colleagues or family

Seven items, each a short symptom phrase. All under 10 words. No "1." numbering — this isn't a sequence.

## Table Snippets (Comparisons)

Tables fit "X vs Y" and pricing queries. Use a real `<table>` element with `<thead>` and `<tbody>`. 3-6 rows, 3-5 columns, short cell content.

### Example 1: "iPhone vs Android"

> ## iPhone vs Android
> | Feature | iPhone | Android |
> |---|---|---|
> | Operating system | iOS (closed source) | Android (open source) |
> | App store | Apple App Store only | Google Play + third-party |
> | Customization | Limited | Extensive |
> | Price range | $429-$1,599 | $100-$1,799 |

4 rows, 3 columns, cells under 10 words. Header column tells the comparison dimension; the two product columns give the facts.

### Example 2: "react vs vue vs angular"

> ## React vs Vue vs Angular
> | Feature | React | Vue | Angular |
> |---|---|---|---|
> | Maintainer | Meta | Evan You / core team | Google |
> | Learning curve | Moderate | Gentle | Steep |
> | Bundle size | ~45KB | ~34KB | ~150KB |
> | Primary use case | SPAs and web apps | Lightweight apps | Enterprise apps |
> | TypeScript support | Optional | Optional | Built-in |

5 rows, 4 columns. Note that 3-way comparisons often get rejected by Google and replaced with an AIO. See `aio-vs-snippet-decision.md` before committing to this format.

### Example 3: "WordPress vs Squarespace"

> ## WordPress vs Squarespace
> | Aspect | WordPress | Squarespace |
> |---|---|---|
> | Cost | Free software, hosting from $5/mo | $16-$52/mo |
> | Ease of use | Steeper learning curve | Beginner-friendly drag-and-drop |
> | Customization | Unlimited via plugins and themes | Limited to built-in options |
> | Maintenance | Self-managed updates | Fully managed |

4 rows, 3 columns. Short cells, clear parallels.

## Critical Rule: First Instance Wins

Google's snippet extractor reads top to bottom and pulls the first substantial instance of the matching format under the matching H2. If your page has the right H2 but the 40-60 word answer is in paragraph 3 (with paragraphs 1-2 being setup), another page with the answer in paragraph 1 will take the snippet. Same rule for lists — if your ordered list is preceded by three paragraphs of context, Google may still extract, but it's less reliable than a clean H2-then-list handoff.

When rewriting, put the answer block immediately after the H2. Setup, context, and commentary all move below the answer, not above it.

# Technique 13: Structured Data & SERP Feature Capture

## What It Is
Using structured data (JSON-LD schema markup) and content formatting to capture SERP features -- featured snippets, FAQ dropdowns, How-To cards, and rich results that increase visibility, CTR, and NavBoost signals.

## Why It Works
SERP features occupy prime real estate above or within organic results. Pages with rich results get 20-40% higher CTR than plain blue links. Higher CTR feeds NavBoost (Google's click-based ranking signal), creating a compound ranking advantage.

Additionally, structured data is how content gets cited by AI Overviews and AI search engines. Without it, your content is less likely to be selected as a source for AI-generated answers.

## SERP Features by Content Type

| Content Type | Target SERP Features | Schema Types |
|-------------|---------------------|-------------|
| Blog post | Featured snippet, PAA | Article, FAQPage |
| How-to guide | HowTo card, Featured snippet | HowTo, FAQPage |
| Product page | Product rich result, Review stars | Product, Review, Offer |
| FAQ page | AI Overview citation, PAA | FAQPage |
| Comparison | Comparison table, Review stars | Article, Product, Review |
| Recipe | Recipe card | Recipe |
| Event | Event listing | Event |
| Local business | Local pack, Knowledge panel | LocalBusiness |

## Step-by-Step Process

### Step 1: Identify Feature Opportunities
1. Use SERP feature detection from to see which features appear for target keywords
2. Search the keyword manually -- what features are currently showing?
3. Note: featured snippets, PAA boxes, HowTo cards, video carousels
4. Check if any competitor has a SERP feature you could capture

### Step 2: Content Formatting for Featured Snippets
5. **Paragraph snippets** (most common):
   - Place a direct answer in 40-60 words immediately after an H2 matching the query
   - Start with the definition/answer, then expand
   - Example: "What is information gain?" -> H2 -> 45-word direct answer -> detailed explanation
6. **List snippets**:
   - Use numbered or bulleted lists with clear H2 header
   - 5-8 items is optimal
   - Each item should be a clear, scannable point
7. **Table snippets**:
   - Use HTML tables with clear column headers
   - 3-5 columns, 4-10 rows
   - Comparison data is ideal for table snippets

### Step 3: FAQ Schema Implementation
8. Identify 3-5 questions users ask about the topic (use PAA boxes as source)
9. Include each question as an H2 or H3
10. Answer directly in 40-100 words
11. Implement FAQPage schema with each Q&A pair:
   ```json
   {
     "@context": "https://schema.org",
     "@type": "FAQPage",
     "mainEntity": [{
       "@type": "Question",
       "name": "What is information gain in SEO?",
       "acceptedAnswer": {
         "@type": "Answer",
         "text": "Information gain in SEO measures how much new, unique information your content provides compared to other pages on the same topic. Google's patented Information Gain Score rewards pages that add original data, perspectives, or insights not found in competing results."
       }
     }]
   }
   ```

### Step 4: HowTo Schema
12. For tutorial/guide content, implement HowTo schema:
    - Each step as a separate `HowToStep`
    - Include `estimatedCost` and `totalTime` when applicable
    - Add `tool` and `supply` arrays
    - Include images per step when possible

### Step 5: Article/Product Schema
13. **Article schema** (for blog posts, guides):
    - `author` with name and URL
    - `datePublished` and `dateModified`
    - `image` (required for rich results)
    - `headline` matching H1
14. **Product schema** (for product/review pages):
    - `offers` with price, availability, currency
    - `review` with `ratingValue` and `bestRating`
    - `brand` entity

### Step 6: Validation
15. Test all structured data with Google's Rich Results Test
16. Monitor Search Console for structured data errors
17. Verify rich results appear in SERPs after indexing

## Tips

- **PAA as keyword research**: "People Also Ask" questions are literally Google telling you what related queries users search. Answer them in your content -> PAA placement and AI Overview citations.
- **Featured snippet "sniping"**: Find a keyword where the current featured snippet is mediocre. Write a better, more direct answer in the same format. Google swaps snippets regularly.
- **Don't over-schema**: Adding 50 FAQ items or marking up content that doesn't qualify for a schema type can trigger manual actions. Only mark up content that genuinely matches the schema type.
- **AI Overview sourcing**: Structured data helps your content get selected as a source for AI Overviews. Content with clear schema markup is easier for AI systems to parse and cite.

## Common Mistakes

1. **FAQ schema for non-questions**: Only use FAQPage schema for actual Q&A content. Using it for bullet points or regular content is a misuse.
2. **Missing required fields**: Schema without required fields (like `image` for Article) won't generate rich results
3. **Duplicate schema**: Having both FAQPage and HowTo on the same page can conflict. Choose the primary schema type.
4. **Not monitoring**: Structured data can break (especially after CMS updates). Monitor Search Console regularly.

## GEO (AI Citation) Impact

Structured data directly impacts AI citation rates:
- **FAQ schema has 3.2x higher citation rate** in AI Overviews
- **92% of AI Overview citations** come from top-10 ranking pages -- SERP features help you get there
- **Answer-first formatting** (40-60 word direct answers under H2s) produces +340% AI citations
- Content with clear schema markup is easier for AI systems to parse and cite

Write comprehensive content (for Google) with clear, extractable summary answers at the start of each section (for AI). The overlap between SERP feature optimization and AI citation optimization is ~80%.

## Tools Used
- SERP feature detection -- identify which features appear for target keywords
- keyword search -- find keywords with featured snippet opportunities
- page-level SEO data -- check current SERP feature status for existing pages