Seo Audit

SEO Audit Skill

A systematic local-service SEO audit workflow for producing client-ready proposals. Covers both B2C single-market (consumer-facing, defined geographic area) and B2B multi-state (contract-based, multi-region, sells to businesses) models.

Step 0: Client intake — ask before doing anything

Run this intake before any research, recon, or analysis. If information is already clear from context, skip that question — don't ask what you already know.

Do not proceed past Step 0 until you have: URL, city, and industry at minimum.

Agency name — used on the cover page and closing section of the PDF. Check memory for the user's company name first. If not known, ask: "What's your agency or company name? This will appear on the cover page and closing." Store this as agency_name for use throughout the deliverable.

Core questions (always ask if not provided)

1. URL "What's the URL of the site we're auditing?" → Used for: live recon, brand color extraction, technical audit

2. Business name "What's the business name?" (or infer from URL if obvious) → Used for: deliverable filenames, report cover, competitive framing

3. Industry / service "What service do they provide?" → Used for: business model classification, competitor research, keyword tiers → ⚠️ Do not assume from the URL — confirm. Similar-sounding services can have completely different business models and competitive landscapes.

Location questions (always ask — these shape the entire audit)

4. Primary city / market "What city or metro is their primary market?" → Used for: Tier 1 keywords ([service] [city]), competitor research scope, Google Business Profile advice, LocalBusiness schema coordinates

5. Neighborhoods or districts served "Do they serve specific neighborhoods or areas within that city? Any areas they're known for or want to target?" → Used for: Tier 2 keyword list, location page roadmap, testimonial seeding → If they don't know: pull from testimonials during recon (Step 1) and come back to confirm: "I found [neighborhood names] in your testimonials — should we target these as location pages?"

6. Geographic scope "Is this just [city], or do they serve a wider region — multiple cities, the whole metro, or multiple states?" → Used for: business model classification (B2C local vs. B2B multi-state), number of location pages needed, keyword volume expectations

7. Service radius (B2C only) If B2C: "How far out do they travel? Do they serve surrounding suburbs or just the core city?" → Used for: extending Tier 2 keywords to adjacent suburbs, additional location pages → Suburbs are often less competitive than the city center — high-value targets

Conditional questions (only ask if relevant)

8. Launch status — ask if unclear from context "Is this site currently live, or is this a pre-launch / rebuild situation?" → If pre-launch: staging domain blocking becomes a hard blocker → If rebuild: ask for old URL to check redirect needs

9. Old domain / previous site — ask if it's a rebuild or rebrand "Was there a previous website or domain? What was the URL?" → Used for: 301 redirect audit, link equity preservation, old URL 404 checks

10. Known competitors — ask to save research time "Are there any competitors you're already aware of — either ones you worry about or ones you want to beat?" → If yes: start research from those, fill gaps to 6-8 total → If no: identify from scratch via web search

11. Priority markets (B2B multi-state only) "Which of your service states or cities are highest priority for new business?" → Used for: location page build order, which markets to research competitively first

12. Existing Google Business Profile "Do they have a Google Business Profile set up? Do you know how many reviews they have?" → Used for: GBP recommendations section, review generation strategy, "near me" keyword advice

What changes based on location answers

This is why location questions matter — each answer has downstream consequences:

After intake: classify the business model

Based on answers, classify into one of two branches:

⚠️ When in doubt, confirm with the user before branching. The entire competitive analysis framework differs between the two.

Step 1: Live site recon

Before any analysis, audit what's actually built — not what the brief says exists.

1a. Browser navigation (use Claude in Chrome tools)

• Navigate to the live URL
• Screenshot the homepage, scroll through all sections
• Note every page in the nav — don't trust what the client told you
• Check footer for sitemap-style links — often reveals pages nav misses

1b. Brand color extraction (always do this — needed for deliverables)

Use javascript_tool to extract exact brand colors from the live CSS:

javascriptCopy code
// Resolve oklch/hsl to hex via canvas
function resolveColor(colorStr) {
  const canvas = document.createElement('canvas');
  canvas.width = canvas.height = 1;
  const ctx = canvas.getContext('2d');
  ctx.fillStyle = colorStr;
  ctx.fillRect(0, 0, 1, 1);
  const [r, g, b] = ctx.getImageData(0, 0, 1, 1).data;
  return `#${r.toString(16).padStart(2,'0')}${g.toString(16).padStart(2,'0')}${b.toString(16).padStart(2,'0')}`;
}

// Get computed colors from key elements
const brandEls = document.querySelectorAll('[class*="brand"], [class*="primary"], [class*="accent"], h1, button, a');
const colors = {};
brandEls.forEach(el => {
  const cs = getComputedStyle(el);
  ['color','backgroundColor'].forEach(prop => {
    const val = cs[prop];
    if (val && !val.includes('rgba(0, 0, 0, 0)') && !val.includes('255, 255, 255')) {
      colors[el.className?.toString().slice(0,50)] = { [prop]: resolveColor(val) };
    }
  });
});
JSON.stringify(colors, null, 2);

Record: primary color, secondary color, accent color, background, text color. These are required for the PDF deliverable.

1c. What to document from recon

Capture structured data for every site you visit — client and competitors.

Business identity:

• Business name (confirm against URL — don't assume)
• Address (real or placeholder?)
• Phone number (check tel: href matches displayed number)
• Email address
• Services offered (list exactly what's on the site)
• Cities / areas served (stated or implied)
• Key selling points (what do they lead with?)

Site structure:

• All live pages (URL + title)
• Navigation structure
• Whether service pages are real URLs or homepage anchor links
• Copyright year (outdated = trust signal problem)

Content & trust signals:

• Real job/team photos or stock?
• Testimonials — do they cite locations, names, specifics?
• Any CTAs that link to # (dead links)
• Stat counters or dynamic elements that may not be rendering

For competitor sites specifically:

• What content gaps exist? What questions does this site fail to answer?
• What would a buyer need to know before calling that isn't addressed here?
• Where is their SEO clearly weak (thin pages, no schema, no blog)?

Step 2: Competitive research

See the relevant reference file for industry-specific competitor lists and what to look for.

General framework (applies to both branches):

Find 6-8 direct competitors. Mix of:

• 2-3 national/franchise players (understand the ceiling)
• 3-4 local/regional operators (understand the real fight)

For each competitor, document:

• Name, URL, headquarters, scale
• Core differentiators (1-2 sentences)
• SEO strengths (content depth, location pages, reviews, schema)
• SEO weaknesses (generic content, no pricing, thin local signals)
• Review volume and average rating

Build a feature matrix — columns are competitors, rows are key SEO features:

• Location/city pages
• Blog/content marketing
• Pricing transparency
• Schema markup
• Review volume
• App or tech differentiator
• Eco/social angle

Step 3: Keyword research

Organize into 4 tiers. See references/keywords.md for full framework.

Location inputs from Step 0 directly feed this step:

• Primary city → Tier 1 ([service] [city])
• Neighborhoods named → Tier 2 starting list
• Suburbs / service radius → extend Tier 2 beyond city center
• Testimonials (from Step 1 recon) → additional Tier 2 seeds if neighborhoods weren't known upfront

Target 40-50 keywords total across all tiers. Document for each:

• Search intent
• Relative difficulty (LOW / MEDIUM / HIGH)
• Target page (homepage, service page, location page, blog)

Step 4: Technical SEO audit

Run through every item in references/technical-checklist.md.

Rate each item: PASS / WARN / FAIL

Step 5: Separate hard blockers from quick wins

Keep these strictly separate in the deliverable:

Hard blockers = site should NOT go live until resolved:

• Placeholder contact info (555 numbers, fake addresses)
• Dead CTAs (links to #)
• Stat counters showing 0 with no real data
• Staging domain not blocked from indexing
• Legal pages missing (Privacy Policy, Terms)

Quick wins = high-impact, low-effort improvements ranked by impact ÷ effort. Common examples: title tag rewrites, meta descriptions, schema markup, sitemap fixes, 301 redirects, service page builds, location page builds. Estimate time for each based on what you actually find — don't use preset estimates.

Always flag blockers in RED in the deliverable. Quick wins in numbered priority order.

Step 6: Content strategy

Two outputs:

Page structure roadmap — what pages should exist:

• Core pages (homepage, about, contact, quote/request)
• Service pages (one per service)
• Location pages (one per market — real pages, not anchor links)
• Audience landing pages (relevant audience segments for the business model)
• Blog

Top 10 blog posts — ranked by search opportunity:

• Must target a specific keyword tier
• Must serve a content purpose (pillar, comparison, FAQ, local)
• Must fit the client's actual expertise

Step 7: Strategic opportunities

Identify 3 specific angles where this client can outmaneuver competitors. These must be grounded in what you actually found during research — not generic SEO advice.

The core question for each opportunity: What are competitors' sites missing? What content gaps exist across the competitive set? What does a buyer need before they'll pick up the phone that nobody in this market is providing well? That gap is the opportunity.

Good examples:

• "No competitor has neighborhood-level location pages despite clear demand signals in testimonials"
• "Competitors rank well but have poor review scores — reliability and trust messaging is uncontested in this market"
• "Nobody in this market publishes pricing — a transparent pricing guide would own that search intent"
• "Competitors' service pages don't explain the process — buyers can't visualize what happens after they call"

Bad examples (too generic, don't use):

• "Create great content"
• "Build more backlinks"
• "Improve site speed"

Each opportunity should name the specific gap found, why it matters for this market, and what the client should build or publish to own it."

Step 8: Generate deliverables

One deliverable per audit:

Full PDF audit report

See references/deliverables.md for requirements and creative guidance.

Sections (in order):

1. Cover page (client name, brand colors, "Prepared by [agency_name]")
2. Table of contents
3. Executive summary with stat callout bar
4. Business overview & site analysis
5. Competitor analysis with feature matrix
6. Keyword research tables by tier
7. Technical SEO audit with PASS/WARN/FAIL ratings
8. Content strategy (page roadmap + blog post list)
9. Quick wins ranked list
10. Strategic opportunities
11. Closing / agency positioning

Deliverable naming convention

Copy code
[ClientName]_SEO_Audit.pdf

Goes to /mnt/user-data/outputs/.

Notes on tone and framing

• Never write "create great content" as a recommendation
• Never recommend "build backlinks" without specifics
• Always quantify effort in hours or days
• Always note when a finding is a blocker vs. an improvement
• Agency positioning: we execute the technical work — the audit is the pitch
• The audit closes with [agency_name]'s capability to implement, not a vague "good luck"