# Search Demand Map > > - **Type:** Skill - **Install:** `agentstack add skill-avikbal-dm-claude-seo-skills-search-demand-map` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [avikbal-dm](https://agentstack.voostack.com/s/avikbal-dm) - **Installs:** 0 - **Category:** [Search](https://agentstack.voostack.com/c/search) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [avikbal-dm](https://github.com/avikbal-dm) - **Source:** https://github.com/avikbal-dm/claude-seo-skills/tree/main/plugins/growth-os/skills/search-demand-map - **Website:** https://www.linkedin.com/in/avikbal/ ## Install ```sh agentstack add skill-avikbal-dm-claude-seo-skills-search-demand-map ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # Search Demand Map ## What this does Keyword research usually optimizes for volume. Volume is a trap. A 10,000-search informational term can be worth less than a 90-search bottom-funnel query that buyers type right before they choose a vendor. This skill maps demand by what it is worth to the business, then shows what to chase first. Demand splits into stages. The map respects that, because a top-funnel term and a vendor-comparison term need different pages, different content, and a different place in the queue. ## When to use it Use it to build a keyword or content strategy, to find the gap between what a site ranks for and what its buyers search, or to decide what to write next. Run it before `content-engine`, which turns the map into briefs and a publishing order. ## Data you pull - **Ahrefs**: keyword universe, search volume, difficulty, parent topics, traffic value, the terms competitors rank for that the site does not, and the SERP overview for each target. - **Google Search Console**: the queries the site already earns impressions for, including the page-two terms where a small push captures real clicks, and the impression-to-click gap that flags intent or title mismatch. - **Lead or CRM data**: which queries and pages actually produce qualified leads, so the scoring rewards terms that convert rather than terms that merely attract. ## Method ### 1. Seed from the business, not the tool Start with the buyer's problems and the words they use, drawn from sales calls, the offer, and won deals. Tools expand the seed. They should not set it, or you inherit whatever a competitor happened to target. ### 2. Build the universe and dedupe by intent Pull the full keyword set, then group terms that share a SERP, because Google treats them as one query. Ranking for the cluster head usually pulls the variants along. Do not write a separate page for each near-duplicate. ### 3. Classify intent and stage Label every cluster by intent: informational, commercial investigation, transactional, or navigational. Map each to a buying stage. This decides the page type and where it sits in the funnel. ### 4. Read the SERP, not just the metric For each target, look at what actually ranks. The SERP tells you the page type Google rewards and which features are in play. If the SERP is all comparison pages and yours is a blog post, difficulty score is irrelevant. You have the wrong format. ### 5. Score by lead value Rank opportunities on four things: commercial intent, position in the buying stage, the gap between current and reachable rank, and the lead value the data shows for similar pages. A reachable page-two commercial term beats a number-one informational term most of the time. ### 6. Shape the clusters into hubs Group the winning clusters into hub-and-spoke structures: a pillar for the broad topic, spokes for the specific queries, and an internal link plan that concentrates authority on the commercial pages. ## Output A demand map: clusters ranked by lead value, each with its intent, stage, target page type, the SERP features in play, and a clear first, next, later order. Flag the quick wins where the site already ranks on page two for a commercial term. ## Guardrails - Volume is a vanity metric until you weight it by intent and lead value. - One cluster, one page. Resist a page per keyword. - Let the live SERP override the difficulty score when they disagree. - Keep brand and non-brand separate so brand demand does not flatter the plan. ## Hands off to `content-engine` to brief and sequence the work, `structured-data-engine` where a SERP feature is worth winning, `ai-search-visibility` for the queries that resolve inside AI answers, and `search-experience` to confirm the page type matches what the SERP rewards. --- ## About the author Avik Bal is a B2B digital marketing practitioner specializing in web architecture, SEO, content strategy, and marketing analytics. He has helped enterprise software, fintech, and technology companies drive growth through scalable digital marketing programs. Avik is the author of *CITED: The Growth Operating System for AI Search*. Part of the Growth Operating System. https://github.com/avikbal-dm/claude-seo-skills ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [avikbal-dm](https://github.com/avikbal-dm) - **Source:** [avikbal-dm/claude-seo-skills](https://github.com/avikbal-dm/claude-seo-skills) - **License:** MIT - **Homepage:** https://www.linkedin.com/in/avikbal/ Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** no - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/skill-avikbal-dm-claude-seo-skills-search-demand-map - Seller: https://agentstack.voostack.com/s/avikbal-dm - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.