Web Content

Process

1. Gather Content Brief

Ask the user:

Required:

• "What topic or page do you need content for?"
• "Who is the target audience?"

Optional (offer but don't require):

• "What's the primary goal? (rank in AI answers, inform, convert, educate)"
• "Any specific keywords or queries this should answer?"
• "What entities/brands/tools/names should be mentioned?"
• "Approximate word count? (default: 1200-1800 words)"
• "What language? (Spanish, English)" — Default to Spanish if not specified.
• "Any existing page or competitor URL to reference?"

2. Generate Content

Produce the full page content following every principle below. Present the output in clean markdown.

Writing Principles (Apply All of These)

Principle 1: Ski Ramp Structure

Invert the traditional narrative arc. Do NOT build to a climax. Instead:

1. Opening section (first 30%): Lead with your strongest claims, key definitions, primary entities, and the direct answer to the page's core question. This is where 44% of AI citations come from.
2. Middle section (30-70%): Expand with supporting evidence, comparisons, examples, and secondary entities. Each paragraph must add new information.
3. Final section (70-100%): Reinforce key points, provide implementation details or next steps. Do NOT introduce new critical information here.

❌ Wrong (narrative arc):
"In today's fast-paced world... [500 words of context] ...the answer is X."

✅ Right (ski ramp):
"X is [direct definition]. It works by [mechanism]. Key tools include A, B, and C.
[Then expand with evidence and detail]"

Principle 2: Definitive Language

Use direct, declarative statements. Prefer subject-predicate-object structures.

❌ "In many ways, automation could be considered a key part of..."
✅ "Demo automation is the process of using software to create interactive product demos."

• Use "is", "are", "refers to" for definitions
• Lead with the subject, not with qualifiers
• One claim per sentence when defining concepts
• Avoid hedge words: "might", "could", "perhaps", "in some cases"

Principle 3: Question-Answer Headers

Format major sections as literal questions in H2/H3 tags, with answers starting immediately in the first paragraph.

❌ <h2>The History of SEO</h2>
   <p>It began in the early 90s when...</p>

✅ <h2>When did SEO start?</h2>
   <p>SEO started in the mid-1990s when webmasters began optimizing...</p>

• Entity echoing: Repeat key terms from the header in the opening sentence
• 78.4% of Q&A citations come from heading-paragraph pairs
• Use questions your audience actually asks (think "People Also Ask")

Principle 4: Entity Richness (20%+ Density)

Pack content with specific proper nouns: brand names, tool names, people, places, frameworks.

❌ "There are many good tools for this task."
✅ "Top tools include Salesforce, HubSpot, and Pipedrive."

• Aim for 20%+ entity density in key paragraphs
• Name at least 3 specific entities per major claim
• Entities act as "anchors" that lower AI uncertainty
• Generic advice = high perplexity = gets skipped

Principle 5: Analyst Voice (Balanced Sentiment ~0.47)

Balance facts with analysis. Neither dry encyclopedia nor hype blog.

❌ Too dry (0.1): "The iPhone 15 uses an A16 chip and has a 48MP camera."
❌ Too hype (0.9): "The INCREDIBLE iPhone 15 is ABSOLUTELY the BEST phone ever!!!"
✅ Analyst voice (0.47): "The iPhone 15 features the A16 chip. Its low-light photography performance makes it a strong choice for content creators."

• State the fact, then explain how it applies or why it matters
• Use analytical phrases: "makes it suitable for", "positions it as", "which means that"
• Avoid superlatives without evidence
• Avoid purely factual lists without interpretation

Principle 6: Business-Grade Readability (Flesch-Kincaid ~16)

Write at college level — clear but not dumbed down. Avoid academic density.

❌ Grade 19+ (academic): "The multifaceted paradigmatic implications of cross-functional synergistic methodologies necessitate comprehensive evaluative frameworks."
✅ Grade 16 (business): "Cross-functional teams need clear evaluation frameworks to measure their impact."

• Short-to-moderate sentences (15-25 words average)
• Subject-verb-object structure
• One idea per sentence
• Technical terms are fine — jargon chains are not
• Break complex ideas into multiple clear sentences

Principle 7: Information Density Per Paragraph

Every paragraph must earn its place by adding new information.

• No "throat-clearing" paragraphs ("In today's world...", "As we all know...")
• Each paragraph should contain at least one entity or specific data point
• Middle sentences matter most (53% of citations) — put your strongest info-gain there
• If a paragraph restates what was already said, cut it

Content Structure Template

When generating content, follow this structure:

# [Page Title — Clear, Entity-Rich]

[Opening paragraph: Direct answer to the page's core question. Include primary definition using "is/are". Name 2-3 key entities. This paragraph is your #1 citation target.]

[Second paragraph: Expand on the answer with a specific data point or comparison. Add context for WHY this matters.]

## [Question-format H2]?

[Answer starting with entity echo from the header. Definitive language.]

[Supporting detail with named entities and analyst-voice interpretation.]

## [Question-format H2]?

[Direct answer + entities.]

[Evidence, examples, or comparisons with specific names/numbers.]

## [Question-format H2]?

[Continue pattern...]

## [Practical/Implementation H2 — can be statement format]

[Actionable content, tools, steps. High entity density.]

## [Closing/Summary H2]

[Reinforce key points. No new critical information. Optional CTA.]

Quality Checklist (Verify Before Delivering)

Before presenting the final content, verify:

• Ski ramp: First 30% contains the primary claim, core definition, and key entities
• Definitive language: Opening paragraphs use "is/are/refers to" (no hedge words)
• Q&A headers: At least 60% of H2s are phrased as questions
• Entity echoing: First sentence after each H2 repeats key terms from the header
• Entity density: Key paragraphs contain 3+ named entities (brands, tools, people, places)
• Analyst voice: Each major claim has both a fact AND an interpretation/application
• Readability: Sentences average 15-25 words; no jargon chains
• No throat-clearing: First paragraph goes straight to the answer (no "In today's world...")
• Information gain: Every paragraph adds something new (no restatement filler)
• Footer discipline: Last 10% reinforces, does not introduce new claims

Output Format

Present the content in clean markdown, ready to paste into a CMS. After the content, include:

Content Scorecard

Entities Used

List all proper nouns (brands, tools, people, places) mentioned in the content.

Target Queries

List 3-5 specific questions this content is positioned to answer in AI search results.

Offer Refinement

After delivering, say: "¿Quieres que ajuste algo? Puedo aumentar la densidad de entidades, cambiar el tono, agregar más secciones Q&A, o adaptar para un público diferente."

Research Source

These principles are based on Kevin Indig's analysis of 1.2 million ChatGPT search results with 18,012 verified citations (Growth Memo, 2026). The study found a statistically significant "ski ramp" citation pattern (P-Value = 0.0) and identified five linguistic characteristics that distinguish cited content from skipped content.