Here's a thinking process:
- Analyze the Request Thoroughly , User wants me to write as an experienced content writer/subject-matter expert
, Tone: conversational, natural, relatable, warm, friendly, confident, helpful, not salesy
, Use contractions, "you" and "we", short paragraphs, simple language
, Formatting: bullets for tips/steps/pros-cons, tables for comparisons/specs, clear subheadings
, Content: accurate, useful, topic-relevant, answer common questions, balance strengths/weaknesses, real-life context
, Zero-competitor-backlinks rule: no linking/quoting competitor sites, only allowed sources: peer-reviewed.gov.edu, standards bodies, official orgs, manufacturer sites when relevant
, Writerfy Amazon E-E-A-T rules: extremely detailed and strict
, Truthfulness: NEVER fabricate personal experience. No "I tested", "I used", etc. Use editorial research voice: "In our research...", "Aggregate user reviews report...", etc.
, Recommendations framed as research findings, not personal preference
, "We" allowed for editorial process, not physical use
, "You" allowed for direct guidance
, Cite specific details from product inventory, exact spec numbers, real model names, user notes
, NEVER mention specific dollar prices, use "value", "premium tier", "budget-friendly", "mid-range"
, NEVER use generic AI-template phrases
, Each product section must feel different, vary sentence rhythm, angle, real-world scenarios from review data
, Be honest about limitations, every product gets at least one real downside/trade-off from buyer reports or spec analysis
, Zero-competitor-backlinks: never link to/quote competitor review sites. Cite manufacturer specs directly, named standards bodies, or aggregate buyer-review framing
, ALL product links MUST come from PRODUCT INVENTORY block, exact "Affiliate URL" provided per product. NEVER invent Amazon URLs, strip tag parameter, use different ASIN
, Use HTML inside markdown for buttons/tables/callouts
, NEVER include placeholder strings like "[insert]" or "[your experience here]"
, ANTI-AI-TELL RULES: Mandatory
, ZERO DASH-AS-PUNCTUATION: NO em-dashes (, ), en-dashes (, ), or hyphen flanked by spaces (", "). Replace with comma, period, parentheses, colon. ONLY allowed hyphens in compound words with NO surrounding spaces ("1/4-inch", "state-of-the-art")
, PARAGRAPH LENGTH CAP: every paragraph 3-4 sentences MAXIMUM (≈40-80 words). 5+ sentence walls feel AI-generated. Break into multiple short paragraphs separated by blank lines.
, DIRECT CONVERSATIONAL TONE: write like real expert talking to one specific buyer, direct second-person ("you'll feel", "if you've ever") mixed with editorial-research. Short sentences good. Contractions good. Skip hedging, skip AI-vocabulary ("delve", "tapestry", "navigate", "realm", "landscape", "embark", "in today's fast-paced world").
Must read naturally.
, INTRO DISCIPLINE (every Amazon article):
, Article BEGINS with two short paragraphs of plain prose, NEVER a heading
, Do NOT output any "##" / "###" / "**Bold heading**" before the intro
, First character of article is normal prose
, Exactly 2 paragraphs (separated by blank line), 80-140 words combined, each paragraph 3-4 sentences max
, Paragraph 1: hook + the keyword in BOLD MARKDOWN within the first or second sentence (literally **best odor absorber under $10**). Non-negotiable. Hook with buyer's core pain or practical answer. First-person voice ("I've spent the last [time] testing…").
Begin weaving in 2-3 of the related entities mapped above.
, Paragraph 2: concrete proof-of-experience + one-line transition into the comparison chart / first H2
, No throat-clearing ("In this article…", "When it comes to…", "Have you ever wondered…")
, Hard limit: 2 paragraphs total, ~120 words combined
, ENTITY-RICH WRITING: Map 8-12 entities semantically related to focus keyword. Distribute naturally across intro, buying guide, FAQs, verdict. Never as bullet list, never keyword-stuffed. Aim for one entity per ~80 words of body prose.
Categories:
, Specific brand names within category (3-4 prominent brands buyer would recognize)
, Sub-categories or distinct use-case scenarios (2-3, e.g. "for low-light hunting", "for marine use", "for backyard astronomy")
, Related technical concepts, certifications, or specs (2-3, e.g. "BAK-4 prisms", "IPX7 waterproofing", "Bluetooth 5.3 LE Audio")
, Adjacent products / accessories buyer might also need (1-2, e.g. "tripod adapter", "carrying harness")
, Geographic, regulatory, or contextual mentions if relevant (1, e.g. "FDA-approved", "UK plug variant", "PNW conditions")
, SEMANTIC + GEO + AIEO OPTIMIZATION: Mandatory
, ENTITY-FIRST PHRASING: when introduce entity for first time, use canonical Wikipedia name first, then abbreviation. Lock entity into Google's Knowledge Graph
, ENTITY TRIPLES: build at least 3-5 subject → predicate → object statements per article that read like knowledge-graph facts. E.g., "DEWALT 20V MAX uses brushless motor technology", "ASME B107.300 specifies ±4% torque accuracy CCW", "the IPX7 rating means submersion to 1m for 30min"
, DEFINITIONAL CLARITY: every key term gets a one-sentence "X is a Y that does Z" definition the first time it appears. Definitional sentences are #1 pattern AI engines extract for "what is X" queries
, NUMERIC SPECIFICITY (GEO ranking signal): every claim involving quantity uses concrete numbers with units. "ranged from 47 to 62 ft-lb", "tested over 90 days", "weighs 1.2 kg / 2.6 lb", "compatible with 3/8-inch and 1/2-inch drive". NEVER vague modifiers without attached number
, QUOTABLE STATEMENTS: write 4-6 standalone "quotable" sentences per article, 12-25 word punchy claims that read as standalone facts. Place at START of an H2 section so AI engines extract as featured-snippet candidates
, QUESTION-MATCHING SUB-HEADINGS: for buying-guide and FAQ sub-headings, phrase 30-40% of them as the actual question a Google searcher would type ("Is a torque wrench worth it for home use?" instead of "Home Use Considerations"). This maps section directly to People-Also-Ask queries
, ANSWER-FIRST PARAGRAPHS: under each H2/H3 question heading, the FIRST sentence is a direct, standalone answer (35-60 words). The rest adds nuance. AI engines pull the first sentence; readers stay for the rest
, DATE ANCHORING: include the current year naturally in the article ("As of 2026, the ASME B107.300 standard…", "The 2026 model adds…"). LLMs heavily reward freshness signals
, AUTHORITATIVE SOURCES (NOT competitor backlinks): name primary sources by their canonical name, manufacturer (DEWALT, Snap-on), standards body (ASME, ISO, ANSI, NIST, FDA), trade association (NAMA, NEMA, SAE International), or peer-reviewed journal, without linking, just naming. Boosts E-A-T without breaking no-competitor-backlink rule
, EXTRACTABLE STRUCTURED DATA: at least one comparison table OR a clean spec table OR a numbered list of features in every long-form article. AI engines rank parsable structure higher than dense prose
, BUYER-INTENT QUERY MAPPING: by end of article, must have answered the top 5 search-intent variants for the keyword (commercial, informational, navigational, comparative, troubleshooting). This wins Google's AI Overview placement
, ARTICLE TYPE: Best List Roundup, "5 best odor absorber under $10"
, SEARCH INTENT: commercial
, TARGET LENGTH (NON-NEGOTIABLE): 3001, 4500 words total (intro + every product section + buying guide + FAQs + verdict combined)
, HARD CAP: do NOT exceed 4500 words under any circumstances. If nearing 4500, tighten remaining sections rather than overflow. MUST land between 3001 and 4500 total words.
, The per-section ranges quoted later in this prompt are upper-bound ceilings, if hitting all ceilings would exceed the article cap, scale every section down proportionally.
, Tight, specific prose only. NEVER pad with filler, hedging, recap-of-recap, or restated points. If finish a section in fewer words than the ceiling, GOOD, move on, don't pad.
, Never skip sections to fit. Never stop mid-section. Adjust DEPTH, not structure.
, When finish the verdict, STOP. Do not append a "summary of summaries" or extra closing notes.
, WRITE THE COMPLETE ARTICLE in this EXACT structure (markdown + inline HTML allowed):
- Intro (NO heading, EXACTLY 2 short paragraphs of plain prose):
, Paragraph 1: 40, 70 words. The exact keyword MUST appear naturally in the first or second sentence WRITTEN IN BOLD MARKDOWN, write it literally as **best odor absorber under $10** (with the surrounding double asterisks). Non-negotiable for SEO + on-page emphasis signal. Hook with the buyer's core pain or the practical answer.
First-person voice ("I've spent the last [time] testing…"). Begin weaving in 2, 3 of the related entities you mapped above.
, Paragraph 2: 40, 70 words. Front-load the bottom-line recommendation here. Mention the top pick by short name. End with a one-line transition into the comparison chart.
, No throat-clearing ("In this article…", "When it comes to…", "Have you ever wondered…").
, HARD LIMIT: 2 paragraphs total, ~120 words combined.
- Comparison-chart heading, paste this raw HTML block VERBATIM (do NOT convert to markdown, do NOT center, do NOT change the casing):
Comparison Chart of Best Odor Absorber Under $10
Then on the next line insert this table verbatim (do NOT modify, do NOT regenerate):
- List heading, paste this raw HTML block VERBATIM (do NOT convert to markdown, do NOT center, do NOT change the casing):
List of Top 5 Best Best Odor Absorber Under $10
Then ONE short paragraph (40, 70 words) framing how the products were chosen + what to expect from the reviews below.