Capstone: Run a Full SEO Audit and Growth Plan

Record the organization, site owners, authorized hosts, business model, real service areas, priority offers, audience/languages, regulated topics, baseline period, analytics properties, Search Console access, CMS/hosting constraints, and stakeholders. State what you cannot access.

Do not crawl, test, or change a system without permission. Remove session tokens, personal data, form submissions, and sensitive query parameters from exports before using AI.

Test critical GA4 events using controlled actions, check that no PII is transmitted, document consent behavior, and reconcile the business’s definition of lead or sale. Review Search Console property coverage, sitemap submission, performance filters, and important manual/security notices where access permits.

Create a measurement-gap table with evidence, impact, owner, fix, and validation step. Measurement defects come before growth conclusions.

Build a keyword and question set from Search Console, customer conversations, onsite search, official terminology, and careful competitor observation. Preserve source, geography, language, and date. Cluster by task and intent rather than word similarity alone.

Map each cluster to an existing page, a justified new page, or “do not target.” Flag cannibalization, unsupported locations, regulated claims, and topics outside the business’s real expertise.

For priority pages, review purpose, evidence, author/reviewer, freshness, title, H1, headings, internal links, descriptive anchor text, image alternatives, structured data eligibility, and conversion path. Separate factual gaps from stylistic preferences.

Recommend consolidation or removal only after checking traffic, links, business purpose, and redirect destination. Do not bulk-delete pages because an AI calls them “thin.”

Combine an authorized crawl with manual fetch/render checks, Search Console, server/configuration evidence, and logs where available. Review status codes, redirects, canonicals, robots controls, index directives, sitemaps, mobile rendering, Core Web Vitals, structured data, internal discovery, duplicate URL patterns, and security/HTTPS basics.

For every issue, save an example URL, evidence source, affected scale, confidence, proposed fix, regression risk, owner, and validation method. Do not promise ranking recovery.

Inventory the strongest existing resources, relevant mentions, referral behavior, partnerships, and legitimate PR opportunities. Review placement relevance and relationship instead of relying on a domain score. Reject paid-followed links, PBNs, mass guest posts, fabricated studies, and undisclosed sponsorship.

Propose one link-worthy asset whose evidence and maintenance owner are clear.

Sequence work in three phases:

• Days 1–30: repair measurement and high-risk technical blockers; validate baselines.
• Days 31–60: improve a controlled priority page set and launch one content or snippet experiment.
• Days 61–90: evaluate evidence, expand only supported patterns, and release the link-worthy asset or focused PR test.

Each backlog item needs business relevance, evidence, confidence, effort, risk, dependency, owner, due date, acceptance test, and rollback. Capacity must fit the client’s actual development and approval speed.