Every format of the same audit, on one page, with nothing gated: PDF, Word, Excel, CSV and Markdown. Pick the one that matches what you are doing and the file builds in your browser as you click.
The audit itself is real — a live site scored across six weighted pillars, with 14 findings, the measurement behind each one, and the checks that passed left visible.
Download the PDF report
The full audit of pearllemon.com — 72/100 across six pillars, 14 findings. No email, no signup; the file is built in your browser when you click.
Same report, other formats
Which format you probably want
The formats are not interchangeable. Each one exists because a different job needs it.
PDF — sending it to a client or a prospect. Final, four pages, survives forwarding.
Word — writing your own audit from this structure. Editable, opens in Google Docs.
Excel — turning findings into a work plan. Four sheets, filter by severity.
CSV — importing findings into a tracker or portal. Findings table only, no formatting.
Markdown — feeding the report into documentation or a language model.
Why nothing here is behind an email form
Gating a sample report is the standard move and it is the reason most of these pages get no engagement: the person searching is comparing five results and will not fill in a form to find out whether yours is any good.
We would rather you download the file, see that the findings carry real measurements, and remember where it came from when you need audits at volume.
What makes this different from the usual sample
Published audit samples are nearly always mock-ups of a tidy site with three cosmetic problems, because nobody publishes a client's real failures and nobody publishes their own. This one is a site in our group, so it includes a security pillar that scores badly and an authoritativeness sub-score of 10 out of 100.
That matters if you are using the download to learn how to write up bad news, which is the hardest part of an audit and the part no template teaches.
Getting the same download for your own domain
Running a free audit produces this identical set of files for any site in about a minute. Paid plans put your agency's branding on the exports and can serve the report from your own domain.
Score breakdown
pearllemon.com scored 72/100 when we audited it on . The overall figure is a weighted average of six pillars, so you can always reconstruct it by hand. The benchmark column is the average across 126 sites we have audited.
3 critical issues, 7 warnings and 4 passed checks. Each one records what was measured, not just a verdict — that evidence line is what makes an audit defensible when a client pushes back.
Critical (3)
No agent interfaces published
AI & E-E-A-T readiness
What we measured:
Agent-interface score 5/100. No /.well-known/agents.json, no WebMCP script tags, no AGENTS.md.
Recommendation:
Publish agents.json declaring your key user flows, add AGENTS.md, and expose WebMCP tools so AI agents can act without scraping the DOM.
Authorship and dates are invisible to machines
AI & E-E-A-T readiness
What we measured:
Authoritativeness score 10/100. No schema.author, no machine-readable <time datetime=…>, 0 authoritative outbound citations.
Recommendation:
Add a visible author byline wired to schema.author, surface a publish/updated date in <time>, and cite two to three .gov/.edu/.org sources.
Security header coverage is incomplete
Security
What we measured:
Security pillar scored 48/100 against the header and transport checks.
Recommendation:
Add Content-Security-Policy, Strict-Transport-Security, X-Content-Type-Options, Referrer-Policy and Permissions-Policy at the edge.
Warning (7)
33 render-blocking scripts and 45 external stylesheets
Performance
What we measured:
111 external resources in total; HTML document weighs 688 KB before assets.
Recommendation:
Defer non-critical JavaScript, combine and inline critical CSS, and drop unused plugin stylesheets. This is the single biggest lever on LCP here.
No images are lazy-loaded
Performance
What we measured:
0% of 58 images use loading="lazy"; only 41% use a modern format.
Recommendation:
Add loading="lazy" below the fold and serve WebP/AVIF for the remaining 59% of images.
16 form inputs have no associated label
Accessibility
What we measured:
16 unlabelled inputs, 14 links with no discernible text, no <main> landmark.
Recommendation:
Associate every input with a <label>, give icon-only links accessible names, and wrap the primary content in <main>.
Only one question-style H2
Content
What we measured:
AI citation score 55/100; one question-format heading across 14 H2s. Answer-first opening paragraph was detected.
Recommendation:
Convert two to three H2s into the questions people actually search ("What is…", "How do I…") — this is what answer engines quote.
Primary call to action sits far below the fold
Content
What we measured:
First CTA appears 7,268 px down the page. Conversion-readiness grade 88/100 otherwise.
Recommendation:
Move a primary CTA into the first viewport.
Inconsistent NAP formatting
Technical SEO
What we measured:
Two different phone-number formats detected across the page and schema.
Recommendation:
Normalise every phone number to a single E.164 format in both copy and LocalBusiness schema.
No web app manifest
Technical SEO
What we measured:
PWA score 3. Favicon and apple-touch-icon present; manifest and theme-color absent.
Recommendation:
Add a manifest.json with icons and a theme colour.
Passed (4)
Crawlability is clean
Technical SEO
What we measured:
robots.txt present, XML sitemap available, canonical declared, all 18 priority AI bots allowed.
Recommendation:
No action needed. Re-check after any hosting or CDN change.
Structured data is well covered
Technical SEO
What we measured:
Structured-grounding score 90/100: Organization, LocalBusiness, Article-class and FAQPage schema all present.
Recommendation:
No action needed. Keep FAQPage in sync with the visible copy.
Trust and social proof are strong
Content
What we measured:
Testimonials, case studies, client logos, trust badges and a Trustpilot profile all detected.
Recommendation:
No action needed.
Image alt coverage is near complete
Accessibility
What we measured:
57 of 58 images carry alt text; a skip link and <nav> landmark are present.
Recommendation:
Fix the single remaining image and this check is clean.
Raw measurements
The underlying numbers the scores were derived from. A report that shows only scores is impossible to check; a report that shows measurements can be re-run by anyone.
Measurements recorded during the audit of pearllemon.com
Metric
Value
Words of body copy
2,780
Heading structure
1 × H1, 14 × H2, 13 × H3
Internal links
134
External links
15
Images
58 (1 missing alt text)
Modern image formats
41% of images
Lazy-loaded images
0% of images
Render-blocking scripts
33
External CSS files
45
Total external resources
111
HTML document size
688 KB
Readability (Flesch)
52 — grade 9.9, "difficult"
AI crawlers allowed
18 of 18 priority bots
How 72/100 compares
Every completed audit in the LemAudit database that carries a full pillar breakdown. Sites were submitted by agencies and prospects, so the sample skews towards small and mid-sized business websites rather than enterprise.
The average site in that sample scores 65/100. Two patterns hold almost universally: accessibility and technical fundamentals are in decent shape, while content depth and machine-readable authorship are not. Content is the weakest pillar overall at 31/100.
Most common worst-issue across 126 audited sites
Most common worst issue
Sites affected
No agent interfaces detected
104 of 126
Very thin content
17 of 126
No WebMCP tools exposed
12 of 126
Duplicate <title> tags
5 of 126
Missing H1 tag
2 of 126
robots.txt blocks all crawlers
2 of 126
Frequently asked questions
Is the download really free with no signup?
Yes. Every file is generated client-side from the data on this page, so downloading sends nothing to us and there is no form in the way.
What formats are available?
PDF, Word (.docx), Excel (.xlsx), CSV and Markdown. The PDF is the client-facing one, Word is editable, Excel is filterable, CSV is for importing and Markdown is for documentation and AI tools.
Can I reuse the report for my own clients?
Yes. Take the Word or Excel version, replace the data with your client's, and delete our footer. The structure is the part worth reusing.
Is there a blank template as well as the filled-in report?
Yes, a blank Markdown template with the same sections and the data stripped out. It is linked from the sample report page.
How recent is the audit?
The audit date is printed on the first page of every export, and we re-run it when the underlying site changes materially rather than leaving a stale file up.
