From strategy and keyword research to production and distribution — at scale, without the chaos.
Content marketing sounds simple until you're actually running it. You're managing a content calendar, juggling SEO briefs, chasing approvals, rewriting drafts that missed the brief, distributing to five channels, and somehow also supposed to be tracking performance and reporting to leadership — all at the same time.
The bottleneck is rarely ideas. It's execution bandwidth. Every piece of content moves through research, strategy, writing, review, distribution, and analysis. When those steps are manual and disconnected, you spend more time on the process than the thinking.
Here's how I mapped my entire content marketing workflow to 10 Claude skills — and what changed when I did.
The most common content mistake is producing before strategising. Content without a clear audience, intent, and distribution plan is just writing. Before you create anything, know: who is this for, what do they search for, what do they already believe, and what do you want them to do next?
A keyword tells you what people type. Search intent tells you what they actually want. "Best CRM software" has transactional intent — the person wants to buy. "What is CRM software" has informational intent — they're learning. Write for the intent, not just the keyword. Google ranks content that satisfies intent, not content that mentions a keyword the most.
Top of funnel (TOFU) — broad, educational content for people who don't know you yet. Blog posts, guides, social content. Middle of funnel (MOFU) — comparison, how-to, and use-case content for people considering solutions. Bottom of funnel (BOFU) — case studies, ROI calculators, demos for people close to deciding. Each stage needs different content and different CTAs.
A piece of content that isn't distributed is just a file on a server. Build your distribution plan before you write — email, LinkedIn, repurposing into social, internal linking to existing pages. If you can't answer "how will people find this?" before publishing, you're not ready to publish.
Content strategy starts with understanding the competitive landscape — not so you can copy it, but so you can identify what's missing. marketing:competitive-brief maps your competitors' content positioning, messaging pillars, and the topics they're hitting hardest. What I care about most is the last part: what none of them are talking about.
If five competitors all have "ultimate guides" to the same topic and not one of them has written about the implementation problems their customers actually face — that's your content angle. You're not competing on volume, you're competing on relevance.
Start with three to five direct competitors and look at what content they rank for using free tools like Ubersuggest or Google's "People Also Ask" boxes. You don't need expensive tools to find gaps — you need systematic observation. Brief Claude with what you find and it will surface the patterns.
Go beyond direct competitors. Look at what content adjacent brands — companies targeting your audience for different products — are ranking for. If a productivity tool is dominating content your audience reads, that's a signal about what your audience cares about, even if that brand isn't a direct competitor. Those topics are often underserved in your category.
Re-run this skill every quarter. Competitors shift their content strategy based on what's working. Catching that shift early — before they've built authority in a new topic — is a real edge.
Before writing a single word, marketing:seo-audit maps your keyword landscape — search volume, difficulty, intent classification, and content gaps relative to what already ranks. The output isn't a keyword list; it's a prioritised content opportunity map.
I use it to build clusters: a pillar page targeting a broad head term, supported by cluster pages targeting long-tail variations. Each cluster page links back to the pillar, building topical authority in Google's eyes. This structure consistently outperforms individual one-off articles.
Keyword difficulty (KD) measures how hard it is to rank for a term. A score of 0–30 is low competition, 30–70 is medium, 70+ is very competitive. If your site is new, target low KD terms first — "best [tool] for [specific use case]" rather than just "[tool]." Win the long tail before you chase high-volume head terms.
Don't just look at search volume. Look at SERP features — if a keyword triggers a featured snippet, you can get position zero with a well-structured answer even if you're not ranking #1. Target keywords where the current top results are weak: thin content, outdated information, or poor UX. Those are the easiest wins.
Ask Claude to classify every keyword by intent before you assign it to content. Writing a transactional piece for an informational keyword (or vice versa) is one of the most common reasons content underperforms despite good writing.
marketing:campaign-plan converts your content goals into a structured, executable strategy. You give it your audience, channels, goals, and timeline — it returns a content calendar with topic priorities, funnel-stage mapping, channel distribution, and success metrics by week.
The plan becomes the document your team — writers, designers, social managers — aligns on before anyone opens a Google Doc. It also gives you the authority to say no when someone asks for a last-minute piece that doesn't fit the strategy.
A realistic content calendar for a small team is one long-form piece per week — not one per day. Consistency beats volume every time. It's better to publish one excellent article weekly than five mediocre ones. Set a cadence you can sustain for six months without burning out.
Build your content plan around business seasonality and sales cycles, not just SEO opportunity. If your pipeline historically dips in Q3, your BOFU content should publish in Q2 so it has time to rank. Content SEO has a lag — what you publish today affects traffic in 3–6 months. Plan accordingly.
Feed your competitive brief (Skill 1) and keyword research (Skill 2) into the campaign plan prompt. When all three inputs are connected, the resulting plan is significantly sharper than if you plan in isolation.
marketing:draft-content writes the content — blog posts, landing page copy, social captions, email newsletters, video scripts. You give it the target keyword, audience, intent, angle, and word count. It returns a structured draft with headers, introduction, body, and CTA already in place.
I don't use it to replace writers. I use it to do the structural heavy lifting — outline, first draft, section copy — so writers can focus on what they're actually good at: voice, nuance, and the specific details that make content feel original. The AI does the scaffolding; the writer does the finishing.
Good content answers a specific question better than anything else on the first page of Google. Before drafting, look at the top 3–5 results for your target keyword and ask: what are they missing? What question do they not fully answer? That gap is where your content adds value.
For high-competition keywords, original data, proprietary research, or first-person expertise signals are increasingly necessary to outrank established content. AI-generated drafts are a starting point — layer in proprietary stats, customer quotes, or unique perspectives that can't be replicated. That's what earns backlinks and differentiates in a sea of similar content.
Ask for three different introduction styles — data-led, question-led, and story-led. The intro is what determines whether someone reads the rest. Test which performs better in your email newsletters before committing to a style for blog posts.
The moment content production scales — multiple writers, multiple channels, multiple formats — voice drift starts. One article sounds formal, another sounds casual. One social post uses industry jargon, another writes for a general audience. The brand starts to feel inconsistent, even when individual pieces are good.
brand-voice:brand-voice-enforcement catches this before it reaches readers. Run your drafts through it and it flags anything that deviates from your defined voice with a specific suggested fix — not a vague note like "this sounds off," but a rewrite of the exact sentence.
Brand voice is how your company sounds in writing: the words you use, the sentence length you prefer, the level of formality, and the personality behind the words. Document it in a one-page style guide — three to five voice attributes with a "we say / we don't say" example for each. That's all Claude needs to enforce it.
Voice drift is most dangerous at the TOFU/BOFU seam. Top-of-funnel content often gets written in a more editorial style; bottom-of-funnel content trends toward sales language. If those two voices sound like different companies, it creates friction at the handoff. Consistency across the funnel is what builds the trust that converts.
Run this skill before sending content to external reviewers, not after. Come to the review with copy that's already passed the voice check and you'll spend time on strategic feedback rather than line-editing style.
marketing:brand-review checks content for unsubstantiated claims, factual inconsistencies, messaging contradictions, and compliance flags. In regulated industries — financial services, healthcare, legal, supplements — this is essential before anything goes live. But even in unregulated spaces, a single inaccurate stat or a claim that doesn't match your product can erode trust fast.
I run every piece through this before it publishes. It takes two minutes and has caught things that would have been embarrassing corrections after the fact.
Common content accuracy issues: statistics cited without a source, product claims that are more ambitious than the product actually delivers, and superlatives ("the only," "the best," "the fastest") that aren't substantiated. Each one is a potential credibility hit if a reader checks. Source every stat and soften every unsubstantiated superlative.
Content compliance extends beyond regulated industries. GDPR and CCPA implications arise whenever you reference user data or cookies in content. ASA and FTC guidelines govern testimonials, endorsements, and sponsored content in the UK and US respectively. If you publish anything that could be read as an endorsement or uses customer results, make sure the attribution and disclosure language is correct.
Paste in the full piece including headline, meta description, and any social copy that will accompany it. Inconsistencies between the headline promise and the body content are one of the most common causes of high bounce rates — catch them here.
Traffic numbers without context are just noise. data:analyze turns your content performance data into decisions. Export your analytics data (Google Analytics, Search Console, or your CMS) as a CSV and ask a specific question: "Which blog posts are driving the most pipeline-stage conversions, not just traffic?" or "Which pages have high traffic but low time-on-page, suggesting a content-intent mismatch?"
The output is a ranked analysis with a recommendation — something actionable, not a description of what you already know from looking at the dashboard.
The metric that matters depends on the content's goal. TOFU content should be measured on organic traffic and new visitor rate. MOFU content on time on page, scroll depth, and return visits. BOFU content on conversion rate and pipeline contribution. Measuring every piece by the same metric gives you misleading data about what's working.
Organic traffic is a lagging indicator — it takes 3–6 months for new content to rank. For faster feedback loops, track click-through rate in Search Console (impressions vs. clicks tells you if your title tag and meta description are compelling) and engagement rate in GA4 (sessions where users actually did something, not just landed and left). These give you signal before the ranking data catches up.
Ask a decision question, not a reporting question. "What's our top content by traffic?" is reporting. "Which pieces of content should we update vs. which should we consolidate based on performance trajectory?" is a decision. The second type gets you something you can act on this week.
A 10% improvement in email open rate compounds significantly across a list of any meaningful size. data:statistical-analysis tells you whether the difference between two headline or CTA variants is statistically significant — or whether you're looking at natural variation that would resolve itself with more data.
The minimum standard for any content A/B test (email subject lines, article headlines, CTA copy) is statistical significance at 95% confidence. That means only a 5% chance the difference you're seeing is random. Without reaching this threshold, you're making content decisions on noise.
The easiest place to start A/B testing content is email subject lines — every major email platform has split testing built in. Test one variable at a time (subject line only, not subject line AND send time). Run the test to your full list or a representative sample of at least 1,000 subscribers per variant before calling a winner.
For article headline testing, use your email newsletter as the lab. Send the same article with two different subject lines to two list segments. The open rate tells you which framing resonated more. Then apply the winning angle to the article's H1, meta title, and social copy. You've tested before you've published and can optimise with data rather than instinct.
Before any content pivot based on test results, run this skill with your numbers. "Version B got a 3% higher CTR" might not be significant. Claude will tell you if you need more data or if the result is solid enough to act on.
Content that drives traffic is only half the job. The other half is capturing that traffic and nurturing it into pipeline. marketing:email-sequence designs the full email nurture flow — timing, branching logic for different engagement signals, subject lines, and complete body copy for every email in the sequence.
I use it to build content upgrade sequences: someone downloads a guide or subscribes via a blog post, they enter a 5–7 email sequence that delivers progressively more valuable content and moves them toward a product touchpoint. Done well, this is where content marketing becomes a measurable revenue driver rather than a traffic metric.
A content upgrade is a piece of bonus content offered within a blog post in exchange for an email address — a checklist, template, or deeper guide related to what they just read. It converts casual readers into subscribers at 3–5x the rate of a generic newsletter sign-up, because the offer is directly relevant to what they came for.
Segment your nurture sequences by the content the subscriber first engaged with — it tells you their interest and funnel stage. Someone who downloaded a "beginner's guide to X" gets a different nurture than someone who read a competitive comparison piece. The more the sequence matches the intent that brought them in, the higher the engagement and conversion rates.
Map the nurture sequence before you write the content piece it's attached to. When you know what you're nurturing people toward, you write the top-of-funnel content to naturally set up that journey rather than treating them as disconnected assets.
Content marketing's biggest internal challenge isn't producing good content — it's proving it works to people who think in pipeline and revenue. marketing:performance-report takes your content performance data and builds a stakeholder-ready report: organic traffic growth, content-attributed pipeline, email list growth, engagement benchmarks, and a prioritised list of what to do next quarter.
The format matters as much as the numbers. Leadership needs a narrative, not a dashboard. This skill gives you both.
Always connect content metrics to business outcomes, not just content metrics. "Blog traffic grew 40%" is a content metric. "Blog traffic grew 40%, generating 120 marketing qualified leads at a $31 CPL" is a business outcome. Frame every report with the business impact and you'll have fewer conversations about whether content is "worth it."
Multi-touch attribution is the honest way to report content's contribution to pipeline. A prospect might read three blog posts, attend a webinar, and then request a demo. First-touch attribution gives content full credit. Last-touch gives it none. A linear or time-decay model gives you a more accurate picture of content's role in the buying journey.
Build a dedicated section for "what we learned." Leaders who see you treating content as a learning system — not just a publishing machine — trust your budget more. Include one thing that worked, one that didn't, and one thing you're testing next.
Content marketing without a system is just a backlog of drafts. With one, it becomes a compounding asset — each piece building on the last, each distribution channel reinforcing the others, each data cycle making the next quarter sharper than the one before.
If you're building a content engine and want to pressure-test your strategy, I'd love to compare notes.
hello@kerrycokelekoglu.com kerrycokelekoglu.com