--- name: content-opportunity-discovery description: > Systematically identify the highest-probability next pieces of content to build, using Search Console near-miss data, keyword-tool expansion, and structural gap analysis against what already exists — rather than guessing at topics. Use when a user asks "what should I build next," "where are my content gaps," or before any content-creation session. This is discovery, not creation — feed the output into whatever process actually builds the page. --- # Content Opportunity Discovery Finds specific, evidence-backed next content to build, ranked by probability of success — rather than brainstorming topics with no signal behind them. --- ## Step 0 — Ask before assuming 1. Is Search Console (or equivalent) access available? This is the single highest-value input — pages the search engine already associates the site with, ranking just outside the good positions, are the fastest path to more traffic. 2. Is there a structured inventory of what content already exists (a content matrix, a sitemap, a spreadsheet)? Structural gap analysis needs this to find *combinations* that are missing, not just individual missing topics. 3. Is a keyword-research tool (Bing Webmaster, or equivalent) available as a secondary source? 4. What does the site's content actually look like when it succeeds — which existing format/angle performs best? New candidates should be evaluated against that pattern, not a generic "good content" standard. 5. Is there an audience/monetization skew worth weighting for (e.g. one segment of traffic is worth more per pageview than another)? If so, factor it into scoring, not just raw volume/probability. **If neither Search Console nor a keyword-research tool is available:** don't stop — lean on Step 2 (structural gap analysis, which only needs a content inventory, not any external tool) as the primary source, and use general web search to gauge whether a candidate topic has real demand instead of a keyword tool's volume number. The output will be less precise without external data, but still directionally useful — say so plainly rather than treating a tool-free run as a full substitute. **On getting the tools connected:** Google Search Console and Bing Webmaster Tools are both free — each just needs site-ownership verification (a DNS record, an HTML file, or a meta tag) and takes a few minutes. If the person running this doesn't have them connected yet, mention it's genuinely worth doing for a much stronger version of this skill going forward, but proceed with what's available now rather than blocking on it. --- ## Step 1 — Search-engine near-miss scan (highest priority) Pull query-level performance data and find queries where the site already ranks in the "near-miss" zone — roughly positions 8–20 — with real impression volume. These are queries the search engine has already decided the site is relevant for; a targeted improvement (a new page, or a substantial rework of an existing thin one) has an unusually high chance of breaking through, compared to targeting a completely cold topic. --- ## Step 2 — Structural gap analysis If the site has any kind of repeatable structure (a matrix of entities × content-types — e.g. "we have a calculator for every X, but no comparison page for several X-vs-Y pairs that both have calculators"), look for missing *combinations*, not just missing individual pages. This is usually a much larger and more reliable source of candidates than blind keyword brainstorming, because each candidate is validated by the fact that its component parts already prove demand individually. --- ## Step 3 — Keyword-tool expansion (secondary, use if Steps 1–2 don't yield enough) If the near-miss scan and structural analysis together don't produce enough candidates, expand into new territory using a keyword-research tool for topic areas not yet covered. Treat this as a fallback source, not the primary one — it lacks the "already proven demand" signal the first two sources have built in. --- ## Step 4 — Verify search intent before finalizing candidates For top candidates, check what's *actually* ranking for the target query right now. Confirm the intent matches what would be built (e.g. don't build a long-form article for a query where every real result is a tool/calculator, or vice versa) — a candidate can look great on volume and near-miss position and still be the wrong content format for that specific query. --- ## Step 5 — Score and rank Score every surviving candidate on: search volume/opportunity size, likelihood of ranking (near- miss status weighs heavily here), and monetization/audience value if that's a relevant factor for this site. Present a ranked shortlist with the reasoning per candidate, not just a raw list. --- ## Output shape A ranked shortlist (roughly 10–15 candidates is usually the useful range), each with: the opportunity source (near-miss / structural gap / keyword expansion), the evidence behind it (current position/impressions, or which existing pieces prove the structural gap), the confirmed search intent, and a score/reasoning. Hand this off to whatever process actually creates content — this skill's job ends at the shortlist.