kamskans SEO Content
🛑 READ FIRST: the line that gets sites deindexed
You can produce content fast. Speed is not the risk. Producing many pages whose primary purpose is to rank rather than to help a reader IS the risk, and Google penalises it regardless of whether a human, an AI, or both wrote it. This is the "scaled content abuse" policy. The job of this skill is to channel your speed into a focused set of genuinely useful pages, never a pile of thin ones. If you follow nothing else, follow the Hard rules and the Pre-publish gate below.
What this skill is for
Turning search demand into durable organic traffic: keyword research → intent mapping → a prioritised content plan → drafting pages that are the best answer for their query → on-page optimisation → internal linking → measuring in Google Search Console and iterating. It assumes the operator will publish on their own site (CMS/static site) and has, or will set up, Google Search Console.
The mental model (why this works)
- Search is pre-existing, intent-rich demand. Someone typing a query already has the problem. You do not interrupt or convince; you just need to be the best answer when they look.
- A ranking page compounds. Unlike an ad (stops paying = traffic stops) or a social post (dead in 48h), a page that ranks earns traffic for years. Content is an asset, not an expense.
- Therefore the only durable strategy is to deserve the ranking. Be the best, most useful, most trustworthy result for a specific intent. Every rule below is downstream of that one idea.
HARD RULES (non-negotiable — violating these can get the site demoted)
These encode Google's spam policies and helpful-content / E-E-A-T guidance. Treat them as constraints, not suggestions.
No scaled content abuse. Never generate many pages that mainly exist to capture rankings and offer little original value. The test Google applies is purpose and value, not authorship or volume. Before generating a batch, you MUST be able to say what unique value each distinct page gives. If the honest answer is "it rephrases what already ranks," do not create it.
People-first, not search-engine-first. Write for a human who has the problem. If you would not send the page to a knowledgeable friend asking that question, it is not ready. Do not write "content about a topic to rank"; write "the answer to this person's question".
Every page must clear the value test. Ask: "What does this give the reader that the current top 3 results do not?" Acceptable answers: original data, first-hand experience, a clearer explanation, a real worked example, a genuinely better structure, an honest opinion, an up-to-date correction. "It's longer" or "it has the keyword more times" are NOT acceptable answers.
Demonstrate Experience (the first E in E-E-A-T). Add what a generic model output lacks: real screenshots, real numbers from the product or a test, concrete step-by-step the author actually performed, specific edge cases. First-hand signal is the single biggest quality lever.
Never fabricate. No invented statistics, fake studies, made-up expert quotes, fake author bios or headshots, fake reviews/testimonials, or citations to sources you did not verify. If you state a fact or number, it must be real and, where load-bearing, linked to its source. Fabrication is both a trust-policy violation and a credibility bomb.
Match intent or don't publish. Each query has an intent (informational / commercial / transactional). Serve THAT intent. Do not shove a hard sell into an informational how-to, and do not write a thin info post for a query where searchers want a comparison or a tool.
No black-hat mechanics. No keyword stuffing, no hidden text, no doorway pages (near-duplicate pages targeting tiny keyword variants), no cloaking, no link schemes / bought links, no auto-spun text, no AI-mass-clone pages differing only by a swapped city/keyword.
Programmatic/templated pages are allowed ONLY with real per-page value. A template is fine when each instance carries genuinely unique, useful data (e.g. a real dataset row, a real integration's real details). It is abuse when instances are interchangeable filler. When in doubt, don't.
Quality over quantity, always. A dozen excellent pages out-rank and out-convert hundreds of thin ones, and carry zero penalty risk. If asked to "write hundreds of blogs," push back and propose a focused set.
AI use is fine; low-value content is not. Google permits AI-assisted content as long as it is helpful and not produced to game rankings. So the rules above are what matter, not the fact that an agent wrote it. Follow the platform's disclosure norms where they apply.
Workflow
1. Research — find the demand
Inputs are usually a keyword export (Ahrefs/Semrush) and/or Google Search Console. For each candidate keyword capture: search volume, difficulty, and intent. Prioritise by:
- Intent x proximity-to-conversion. "best X / X alternative / X vs Y / X pricing" (commercial/transactional) convert hardest — do these first. Then high-intent how-tos your product/topic uniquely answers.
- Winnability now. Bias to lower-difficulty terms you can realistically rank for given current site authority. Don't burn effort on head terms a new site can't win yet; come back to them later.
- The content-gap report (keywords competitors rank for that you don't) is the highest-leverage single input — it's a list of proven demand.
- Emerging clusters. A rising, low-competition topic you can credibly own beats fighting for a saturated one. Plant flags early.
Group keywords into topic clusters: one pillar page per cluster + supporting pages, interlinked. Don't create one page per near-identical keyword (that's doorway territory) — one page serves a cluster of close variants.
2. Pick the page archetype (intent-matched)
- How-to / tutorial (informational): the searcher wants to accomplish a task. Deliver the steps, with real screenshots/examples. End with a soft, relevant CTA, not a hard sell.
- Listicle / "best X for Y" (commercial): the searcher is shopping. Be genuinely useful and honest; include real criteria; it's fine for your product to be on the list if it earns the spot.
- Comparison "X vs Y" / "best X alternative" (commercial, high intent): see the dedicated pattern below.
- Glossary / "what is X" (informational, top-of-funnel): clear definition + context; builds topical authority and internal-link targets.
- Pillar / hub (broad): the cluster's home; links out to every supporting page.
3. Draft — be the best answer
- Lead with the answer. Satisfy the query in the first screen; don't bury it under 600 words of preamble.
- Structure for scanning: clear H1 (one per page), descriptive H2/H3, short paragraphs, lists/tables where they help, a TL;DR for long pieces.
- Original elements are mandatory (see Hard rule 4): at least one of a real screenshot, real data, a worked example, or first-hand insight per page.
- One primary keyword + natural variants. Use it in the title, H1, first ~100 words, and naturally throughout. Never stuff.
- Write the depth the query needs, not a word count. Some queries want 300 words; some want 2,000. Match the SERP and the question.
4. On-page checklist (every page)
- Title tag (~50–60 chars): primary keyword near the front + a reason to click. Distinct from the H1 is fine.
- Meta description (~150–160 chars): accurate, compelling; not a ranking factor but drives click-through.
- URL slug: short, lowercase, hyphenated, keyword-bearing, no dates/IDs.
- Headings: exactly one H1; logical H2/H3 hierarchy.
- Internal links: link new pages to/from related cluster pages with descriptive anchor text. This is how authority flows and how Google finds pages. Every new page should link out to 2–4 relevant pages and be linked from the pillar.
- Images: descriptive filenames + real alt text; compress for speed.
- Structured data where it fits (Article, FAQ, HowTo, Product, Breadcrumb) — only mark up content actually on the page.
- One clear, intent-appropriate CTA.
- Internationalisation/canonical: set a canonical; avoid duplicate content competing with itself.
5. Comparison / "alternative" page pattern (highest intent — handle with care)
These convert best, so they're worth doing well and honestly.
- Be accurate and fair. State competitor strengths truthfully. A hit piece reads as biased and erodes trust (and can draw complaints). Win on facts, not spin.
- Use a real comparison table of features/pricing the reader cares about, kept current.
- Be explicit about who each option is best for, including cases where the competitor wins. Honesty here raises conversion, not lowers it.
- Keep claims verifiable. Don't assert a competitor "doesn't have X" if they do; you'll be wrong publicly.
6. Measure & iterate (Google Search Console is the source of truth)
- Track impressions, clicks, average position, CTR per page/query in GSC (free, exact, your own data). This beats any third-party estimate for your own site.
- Leading indicators (weeks): indexed, impressions starting, position climbing into the top ~20. Lagging (months): top-3 + clicks.
- Improve before you add. A page stuck at positions 5–15 with impressions is usually faster to push up (better answer, more depth, internal links, fresher data) than writing a brand-new page.
- Prune or merge pages that never gain traction or cannibalise each other. Fewer, stronger pages is the healthy direction.
Anti-patterns (do NOT do these)
- Mass-producing near-duplicate pages (city/keyword swaps, spun text).
- Writing to a word count instead of to the question.
- Keyword stuffing or unnatural exact-match repetition.
- Fabricated stats, fake authors, fake reviews, unverified citations.
- Hard-selling on informational queries.
- Targeting one page per tiny keyword variant (doorway pages).
- Chasing head terms a new/low-authority site cannot win yet.
- Publishing without internal links (orphan pages).
- "Hundreds of blogs fast" as a strategy — it's the textbook scaled-content -abuse trigger. Reframe to a focused, excellent set.
Pre-publish gate (run this on EVERY page before shipping)
Ship only if all are true:
- Value test: I can name what this gives the reader that the top 3 results don't.
- Intent match: the page serves the query's actual intent.
- Original element: ≥1 real screenshot / real data / worked example / first-hand insight is present.
- No fabrication: every stat/quote/claim is real and load-bearing ones are sourced.
- On-page basics: title, meta, one H1, clean slug, alt text, internal links in and out.
- Not part of a thin batch: this is not one of N interchangeable pages; it stands on its own.
- A knowledgeable human would find it genuinely useful.
If any box is unchecked, fix it or don't publish. When unsure whether a batch crosses into scaled-content abuse, default to NOT publishing and tell the operator why.
Quick reference: priority order for a fresh site
- Bottom-funnel commercial ("best X", "X alternative", "X vs Y", "X pricing") — closest to conversion, often winnable.
- High-intent how-tos your topic/product uniquely answers (each ending in a soft, relevant CTA).
- Emerging low-competition clusters you can credibly own — plant flags early.
- Top-of-funnel informational / glossary — builds topical authority and internal-link targets; lower direct conversion, plays the long game.
Pair every plan with internal linking and GSC measurement. Optimise winners before adding new pages.