Sample SEO audit report: our own website, audited 7 times

<url><loc>https://brutallyhonest.rbamarketing.com/</loc><lastmod>2026-09-25</lastmod></url>

User-agent: *
Disallow: /thanks

• Found: <title>Brutally Honest Website Reviews — Free SEO &amp; AI-Search Audit</title> (60 chars). The brand takes the first 31 characters, and "SEO" doesn't appear until character 34.
• Why it matters: nobody searches the brand yet. The query-bearing half is the part most likely to be truncated or rewritten.
• Fix: <title>Free SEO &amp; AI Search Audit, Graded A+ to F- | Brutally Honest</title>

• Found: no <link rel="icon"> in the head; /favicon.ico and /favicon.svg both return 404.
• Fix: <link rel="icon" href="/favicon.svg" type="image/svg+xml"><link rel="icon" href="/favicon.ico" sizes="48x48"> (square icon, 48×48 or larger).

• Found: H1: Your website probably deserves a D. We'll tell you exactly why.
• Why it matters: a strong hook, but it contains none of "SEO", "audit" or "AI search". The H1 is the strongest on-page topic label.
• Fix: <h1>Your website probably deserves a D. <span>A free, brutally honest SEO and AI-search audit shows you why.</span></h1>

• Found: 84 checks. Nine areas. Weighted on purpose. / Roughly half of the auditor's instructions are restrictions. / One page. One grade. No sales call. These are clever, but a reader skimming only the headings can't tell what each section answers.
• Fix: see the H2 outline in Rewrites.

• Found: the page is a request form with a report delivered later ("Typical turnaround is less than 72 hours"). Searchers typing "free SEO audit" get instant, no-email tools.
• Fix: target "honest website review", "human SEO audit" and "AI search visibility audit". Say plainly why a human-checked report is worth the wait (fix 7).

• Found: the weighting and hard-cap tables are original. Everything else is claims without evidence. The sample card is CSS ([business name redacted] D+ Score 41/100 Product grid invisible to AI), and there are no audit counts or grade distribution.
• Fix: <p>Of the <strong>N</strong> pages we graded between [month] and [month] 2026, the median grade was <strong>D</strong>; <strong>X%</strong> were missing main content from raw HTML.</p>. Use only real numbers.

• Found: An automated audit of 84 checks, scored out of 100, graded A+ to F-. It never says what gets audited.
• Fix: use the opening paragraph in Rewrites.

• Missing: who it's for, which AI engines are tested, how it differs from instant checkers, and paid-tier prices (both paid tiers say "Contact for Quote"). There is also no real before/after example.

• Found: the footer has NAP ("RBA Marketing | 1002 Boardly Hills | Sevierville, TN 37876 | (865) 328-7990") and an /about link. No named person appears anywhere on the page.
• Fix: add a "Who grades your page" block with name, role, years of experience, photo and a link to /about.

• Found: no <time> element and no visible date. "2026" does not appear in main content.
• Fix: <p>Last updated <time datetime="2026-09-25">September 25, 2026</time></p>

• Found: 0 outbound links in <main>. The only external links are the footer's Facebook, Instagram, YouTube and rbamarketing.com. Unsourced claims include "The crawlers behind ChatGPT, Claude and Perplexity read the page as it first arrives and move straight on", "special 'AI instruction files' that Google openly ignores" and "two types of search markup that Google retired".
• Fix: link each claim to its source: Vercel/MERJ AI-crawler study, Google Search Central AI-features guidance, and Google's FAQ/HowTo deprecation notices.

• Found: the only <img> is the footer logo. The "What Google sees / What ChatGPT sees" panels and the grade card are CSS mock-ups.
• Fix: add a real, redacted report screenshot and a real raw-vs-rendered comparison image, each with descriptive alt text.

• Found: every number describes the method ("84 checks", "20 points", "less than 72 hours"). None is a sourced or original statistic.

• Found: no attributed quotes or testimonials. Fix: add 2–3 named client quotes, with permission.

• Found: H3s like It cannot make things up. and It cannot promise you traffic. start with "It" and only make sense in context.
• Fix: "The audit cannot make things up: anything it couldn't measure is marked NOT VERIFIED."

• Found: "Brutally Honest" appears 0 times in main content. The tier heading says Brutal Honest Website Report, the og:image artwork reads "Brutally Honest Website Report", the footer says "RBA Marketing" and the JSON-LD says "RBA Marketing and Media".
• Fix: one product name ("Brutally Honest Website Reviews") and one company name ("RBA Marketing and Media"), used everywhere.

• Found: no visible date and no dateModified. Fix: 4.6 plus "dateModified":"2026-09-25" on a WebPage node.

• Found: the section "Google waits for your page to load. ChatGPT doesn't." has no link to /ai-search (200), the page on exactly that topic.
• Fix: <a href="/ai-search">why AI crawlers miss JavaScript content</a> inside that section.

• Found: Service + Offer "price":"0" fits the page. But Organization and LocalBusiness are two separate nodes with the same name, no @id, and LocalBusiness has no url.

• Found: no WebSite node. Organization has no logo and no sameAs, although the footer links to https://www.facebook.com/rbamarketing, https://www.instagram.com/rbamarketing and https://www.youtube.com/@rbamarketing.
• Fix (replaces the Organization and LocalBusiness nodes; keep the existing FAQPage node):

  {"@context":"https://schema.org","@graph":[
   {"@type":"LocalBusiness","@id":"https://www.rbamarketing.com/#org","name":"RBA Marketing and Media","url":"https://www.rbamarketing.com","logo":"<URL of a logo readable on white>","telephone":"+1-865-328-7990","address":{"@type":"PostalAddress","streetAddress":"1002 Boardly Hills","addressLocality":"Sevierville","addressRegion":"TN","postalCode":"37876","addressCountry":"US"},"sameAs":["https://www.facebook.com/rbamarketing","https://www.instagram.com/rbamarketing","https://www.youtube.com/@rbamarketing"]},
   {"@type":"WebSite","@id":"https://brutallyhonest.rbamarketing.com/#website","name":"Brutally Honest Website Reviews","url":"https://brutallyhonest.rbamarketing.com/","publisher":{"@id":"https://www.rbamarketing.com/#org"}},
   {"@type":"WebPage","@id":"https://brutallyhonest.rbamarketing.com/#webpage","url":"https://brutallyhonest.rbamarketing.com/","isPartOf":{"@id":"https://brutallyhonest.rbamarketing.com/#website"},"dateModified":"2026-09-25"},
   {"@type":"Service","name":"Brutally Honest Website Reviews","serviceType":"SEO and AI-search website audit","provider":{"@id":"https://www.rbamarketing.com/#org"},"offers":{"@type":"Offer","price":"0","priceCurrency":"USD"}}
  ]}

• Found: https://assets.cdn.filesafe.space/jr5tnGZ1CNwntuZjbMrl/media/6aae921ade8ed1c29f6ddbfe.png returns 200, image/png, 1672×940 (1.78:1) and 2,527,439 bytes. The artwork says "Brutally Honest Website Report", which is a different product name.
• Fix: re-export at 1200×630, WebP/JPEG under 300 KB, with the site's real name.

• Fix: <meta property="og:image:width" content="1200"/><meta property="og:image:height" content="630"/><meta property="og:image:alt" content="Brutally Honest Website Reviews: free SEO and AI-search audit graded A+ to F-"/>

• Fix: <meta property="og:site_name" content="Brutally Honest Website Reviews"/><meta property="og:locale" content="en_US"/>

• Found (Playwright, Pixel 5 at 393 px wide): the layout is responsive, scrollWidth 393 = viewport 393 (no horizontal scroll), no visible text under 12 px, and mobile and desktop show the same 1,362 visible words. But the hero URL field, the page's primary conversion control, renders at 361×23 px. The input has class="h-12 flex-1 …". In the mobile flex-direction: column wrapper, flex-1 sets flex-basis: 0%, and that overrides h-12 (48 px). In the screenshot the placeholder https://yoursite.com/the-page-you-care-about touches both borders. That is below the 24 px WCAG 2.2 minimum target size and half the 48 px Google recommends. The footer links are also 16 px tall (What We Check, Pricing, Terms, etc.).
• Why it matters: most small-business owners will open this page on a phone, and the one field they have to tap is a sliver.
• Fix: apply flex-1 only from the breakpoint where the wrapper becomes a row, and set a minimum height:

  <input id="page-url-hero" name="pageUrl" type="url" required
    class="h-12 min-h-12 w-full sm:flex-1 rounded-md border-2 …">

  Add py-2 or min-h-6 spacing to the footer links.

• Found: lab LCP 3.2 s, FCP 3.2 s, Speed Index 4.6 s, TBT 0 ms, performance score 85. LCP equals FCP, so the delay is render-blocking CSS and fonts, not JavaScript.

• Found: two blocking stylesheets, /assets/styles-ASJNGGOi.css and https://fonts.googleapis.com/css2?family=Roboto:wght@300;400;500;600;700;900&family=Roboto+Slab:wght@400;700&family=Permanent+Marker&display=swap. The rendered load then fetched 5 WOFF2 files from fonts.gstatic.com (37,520 + 36,292 + 34,160 + 29,296 + 16,692 bytes = 153,960 bytes of fonts). Total transfer for the initial view is only ~151 KB (the fonts come from cache after the first hit), so page weight is fine and the problem is the cross-origin chain.
• Fix: self-host the fonts, keep only the weights in use, inline critical CSS, and add <link rel="preload" as="font" type="font/woff2" href="/fonts/roboto-900.woff2" crossorigin>.

• Found: Schedule Your Scan (×2) → /pricing. The anchor promises scheduling and delivers a price list.
• Fix: relabel as "See full pricing" or link to /request-report?pageUrl=.

• Found: /request-report (the primary CTA, linked 3 times) returns 307 → https://brutallyhonest.rbamarketing.com/request-report?pageUrl=. None of the other sampled links returned 4xx or 5xx.

• Found, duplicate IDs: both lead forms use <input id="page-url" …> and <input id="company_website_hp" …> with <label for="page-url" class="sr-only">Page URL to grade</label>. Because the IDs repeat, both labels bind to the first input. In the rendered DOM the second pageUrl input has labels: [], so a screen reader or browser agent sees an unnamed URL field.
• Found, contrast (computed in the rendered DOM): the red accent oklch(0.53 0.24 27) fails WCAG AA 4.5:1 on small text. Free · one page · no sales call (14 px) is 3.53:1, and the weighting-table point values 20, 15, 10 (16 px) are 3.53:1. The grey captions Page audit, Sample — critical issues only shown, What Google sees and What ChatGPT sees (12 px, oklab(0.34 … / 0.6)) are 3.35–3.54:1. If the page… and The grade cannot exceed (12 px, 70% opacity) are 4.38:1.
• Fix: give each form instance its own IDs (page-url-hero / page-url-footer, hp-hero / hp-footer) with matching for= values. Darken the red used on text to about oklch(0.45 0.2 27) (or keep the current red for fills only). Raise the caption opacity from /0.6 and /0.7 to at least /0.85.
• Passed within 9.7: the skip link is the first Tab stop (<a href="#main" class="sr-only focus:not-sr-only …">); there are 0 unnamed buttons; the honeypot sits off-screen at left: -9878px with tabindex="-1".