Anatomy of a Modern SERP

The phrase “ten blue links” is a museum piece. A real 2026 Google SERP for a competitive query routinely contains 14+ feature blocks before the first organic listing — AI Overview, People Also Ask, knowledge panel, local pack, video carousel, shopping carousel, sitelinks, Top Stories, related searches, and finally the classic results. Knowing which features your query triggers, and which you can win, is the difference between an SEO who ships and an SEO who optimizes invisible pages.

TL;DR

• The SERP is a feature mosaic. Each query triggers a different combination of features. Your job is to identify which features are present, which you currently win, and which you can realistically capture given your authority.
• AI Overviews are the new featured snippet — and bigger. As of April 2026 they appear on ~47% of US informational queries (Semrush AI Overviews Tracker). Pages cited in them earn a brand mention even when the click rate is reduced.
• Pixel position matters more than rank position. Ranking #1 below an AI Overview, four PAA boxes, and a video carousel can mean your listing is 1,400 pixels below the fold. Pixel-position tracking is now standard in Ahrefs, Semrush, and STAT.

The mental model

A modern SERP is like the front page of a newspaper, not a phone book. The phone book listed everyone alphabetically. The newspaper has a banner story (AI Overview), a side panel (knowledge panel), a photo carousel (image pack), a stock ticker (shopping widget), the editorial column (featured snippet), reader letters (People Also Ask), the local-news box (local pack), a video promo (video carousel), and only in the lower half does it get to the actual classified ads — and even those are interrupted by sidebars and stickies.

The implication: you are not optimizing for “rank one.” You are optimizing for a specific feature on a specific query. The work to win an AI Overview citation is different from the work to win a featured snippet, which is different from the work to win the local pack, which is different from the work to win a Discover impression. You research the SERP first, then decide which battle to fight.

The other implication: classic SERPs barely exist for the queries that matter. Branded navigational queries still look like 2010 — your homepage, sitelinks, knowledge panel, done. But any commercial-investigation query (best running shoes, notion vs obsidian, crm for small business) is now a feature collage where the top 4 features eat 60-70% of the visual real estate.

Deep dive: the 2026 reality

The current major SERP features and what triggers them:

Feature	Trigger pattern	Win condition
AI Overview	Informational + commercial queries; ~47% of US queries (Semrush, Apr 2026)	Cited as a source — clean structure, schema, original data, top-10 organic position helps
AI Mode	Conversational follow-ups via the AI Mode tab; full Gemini-driven dialog	Coverage across query fan-out, entity completeness
Featured snippet	”What is”, “how to”, definition queries (~10-12% of SERPs)	Concise direct answer in HTML, ~40-60 words, paragraph or list format
People Also Ask	Most informational queries (~84% per Mozcast)	Distinct H2/H3 answering each question with 1-3 sentence response
Knowledge panel	Branded entity, person, place, organization queries	Wikipedia + Wikidata entity, schema.org Organization with sameAs links
Local pack	Local-intent queries (“near me”, city + service)	Google Business Profile, NAP consistency, local citations, reviews
Image pack	Visual-intent queries (recipes, products, tutorials)	Image schema, alt text, lossless compression, descriptive filenames
Video carousel	”How to”, entertainment, product-demo queries	YouTube uploads, descriptive titles, chapter timestamps
Sitelinks	Branded queries	Strong site architecture, internal linking, search intent for brand
FAQ rich result	Pages with FAQPage schema (Aug 2023 — most non-gov/edu sites lost eligibility)	Now limited to gov/edu/health authority sites
Top Stories	News-worthy queries within hours-of-publish window	Google News inclusion, NewsArticle schema, fast publishing cadence
Discover	Personalized feed on mobile / Google app	E-E-A-T, large compelling images (≥1200px), recency, user click history
Shopping carousel	Product queries	Merchant Center feed, Product schema, GTIN + offers
Recipe card	Recipe queries	Recipe schema, image, step list, reviews
Job posting	Hiring queries	JobPosting schema, fresh posting date
Event card	Event queries	Event schema, structured location/date
Twitter/X carousel	Trending topics	Real-time integration via Google’s X partnership

The AI Overview specifically deserves attention. It appears as a card at the top of the SERP, summarizes the answer in 100-300 words, and includes 3-7 inline citations with site icons. Citation slots are limited; appearing in one is the new “rank zero.” A Semrush study (March 2026) found AI Overview citation correlates with rank 1-12 in the classical SERP roughly 92% of the time, but the remaining 8% are sites with strong entity signals at lower ranks — Wikipedia, government, well-cited niche authorities.

AI Mode is the chat interface accessed via the AI Mode tab on Google Search (rolled out broadly through 2025-2026). Behind the scenes it issues a fan-out query — a single user question expands to 8-30 underlying searches that the AI synthesizes. To rank in AI Mode you have to cover the cluster, not the single query.

Discover is mobile-only and personalized. It does not respond to a query; it pushes content to users based on interests. Discover impressions can dwarf classic search traffic for news and commentary publishers — it is the closest thing to free social distribution Google offers.

Visualizing it

flowchart TD
  Q["User searches running shoes 2026"] --> AIO["AI Overview (top, 3-7 citations)"]
  Q --> PAA["People Also Ask (4-8 questions)"]
  Q --> SHP["Shopping carousel (top right)"]
  Q --> VID["Video carousel"]
  Q --> ORG["Organic results (rank 1-10)"]
  Q --> RS["Related searches"]
  Q --> AIM["AI Mode tab (conversational)"]
  AIO --> CITE["Cited source = brand mention + click"]
  ORG --> CTR["Classic CTR distribution"]
  AIM --> FAN["Fan-out to 8-30 sub-queries"]

Bad vs. expert

The bad approach

<!-- Targeting "best CRM for small business" with a generic listicle -->
<h1>Best CRM for Small Business</h1>
<p>Looking for the best CRM for your small business? We've reviewed the top CRM options. Here are our picks.</p>

<h2>1. Salesforce</h2>
<p>Salesforce is a popular CRM. It has many features. It is good for businesses.</p>

<h2>2. HubSpot</h2>
<p>HubSpot is another option. It also has features. It is also good.</p>

This fails the AI Overview because there is no extractable answer, no original data, no schema, and no entity clarity. It fails the featured snippet because there is no concise definition or comparison table. It fails People Also Ask because no H2 maps to a real question. It fails the shopping carousel because there is no Product schema. The page is targeting one feature (organic rank) on a SERP where the top 60% of pixels go to features it cannot win.

The expert approach

<h1>Best CRM for Small Business: 14 Tools Tested on 30-Day Pipelines</h1>

<p>For teams of 5-50, the best CRM in 2026 is HubSpot Free for sales-light teams, Pipedrive for outbound-heavy teams, and Close for high-velocity SDR teams. We tested all 14 against the same 30-day sample pipeline.</p>

<table>
  <thead><tr><th>Tool</th><th>Best for</th><th>Starting price</th><th>Setup time</th></tr></thead>
  <tbody>
    <tr><td>HubSpot</td><td>Sales-light</td><td>$0</td><td>2 hours</td></tr>
    <tr><td>Pipedrive</td><td>Outbound</td><td>$15/user</td><td>4 hours</td></tr>
    <tr><td>Close</td><td>High-velocity SDR</td><td>$49/user</td><td>3 hours</td></tr>
  </tbody>
</table>

<h2>Which CRM is best for a 5-person team?</h2>
<p>For a 5-person team, HubSpot Free covers contacts, deals, and email tracking at zero cost. Pipedrive is the next step up at $15 per user when you need pipeline customization.</p>

<h2>What is the cheapest CRM for small business?</h2>
<p>HubSpot Free is the cheapest functional CRM in 2026, with unlimited contacts and 1,000 marketing emails per month at no cost. Zoho Bigin starts at $7 per user per month for the next tier of customization.</p>

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "ItemList",
  "name": "Best CRM for Small Business 2026",
  "itemListElement": [
    {"@type": "ListItem", "position": 1, "item": {"@type": "SoftwareApplication", "name": "HubSpot"}},
    {"@type": "ListItem", "position": 2, "item": {"@type": "SoftwareApplication", "name": "Pipedrive"}},
    {"@type": "ListItem", "position": 3, "item": {"@type": "SoftwareApplication", "name": "Close"}}
  ]
}
</script>

This works because the lead paragraph is an extractable AI Overview citation. The table is a featured snippet candidate. Each H2 is a People Also Ask answer. The schema feeds the shopping/list carousel. The original test data (“30-day sample pipeline”) gives Perplexity and ChatGPT something quotable that is not on competitor pages.

Do this today

1. Run your three highest-revenue queries on Google.com in incognito. Screenshot the full SERP top to bottom. Annotate each feature block.
2. In Ahrefs or Semrush, open Keywords Explorer for the same queries > SERP overview tab. Note the SERP features column — this is your feature inventory.
3. In GSC > Performance > Search results, filter Search Appearance dropdown. Each entry (AI Overview, Web Light, Videos, etc.) shows which features your site already wins.
4. For one query where you rank 8-15 organically, open Semrush AI Overviews Tracker or check Ahrefs SERP overview to see which sources are cited in the AI Overview. Compare your page structure to theirs — typically they have a clearer H1, a 40-60 word answer paragraph, and table or list markup.
5. Add FAQPage schema only if you publish content from a gov, edu, or recognized authority site (everyone else lost FAQ rich result eligibility in August 2023). For all other sites, replace FAQ schema with WebPage + visible H2 questions to win PAA.
6. For local-intent queries, open Google Business Profile > Insights and confirm your profile is complete: hours, photos within 30 days, three or more recent reviews, services list.
7. For news/commentary queries, submit your site to Google Publisher Center at publishercenter.google.com to be considered for Top Stories and Google News surfaces.