We ate our own dog food

BuyerEyes scored our own landing page 68/100. Trust came in at 4.5/10. Here is what the audit found, what we fixed, and where we stand now.

We built a tool that evaluates landing pages. At some point we had to ask the obvious question: what does it say about ours?

On March 12, 2026, we ran a full BuyerEyes audit on buyereyes.ai. 14 agents, 29 sub-scores, 10 buyer personas. The results were uncomfortable enough that we published them here instead of quietly fixing things in the dark.

Methodology note: This audit used the same pipeline available to every BuyerEyes customer — SSR scoring (ρ=0.90), multi-agent debate, saliency heatmaps, and per-persona purchase intent simulation. No special treatment.

The audit results

Six dimensions, one composite score.

68
Overall
out of 100
4.5
Trust
critical
5.0
Visual
below average
9.5
Copy
but too long
7.5
CTA
mobile gap
8.1
Technical
accessibility issues

The overall score is held down almost entirely by Trust (4.5) and Visual (5.0). The primary bottleneck identified by the synthesis agent: Trust & Credibility. The site reads as anonymous. No contact information. No visible social proof. No founder photo. Copy is strong; the page just doesn't feel safe to buy from.

73.8s fullyLoaded time (analytics script)
268ms LCP — actual paint time
122 total requests on load
3,726 words on the homepage

What the audit found

Five agents flagged issues independently. The debate rounds confirmed all of them. Here are the six findings that drove the remediation plan.

Trust · Critical

No contact information anywhere on the page

The automated inventory returned "NONE FOUND" for header and footer contact info. No email, no address, no support link. For a B2B service asking customers to pay $49–$399, this signals zero accountability.

Trust · Critical

No visible social proof or testimonials

The audit found no testimonials, case studies, or third-party review badges on the homepage. For a B2B service asking $49–$399, the absence of any social proof makes the purchase feel risky. This was flagged as the single highest-leverage fix available.

CTA · High

3 competing primary CTAs above the fold

Three distinct conversion actions competed for attention before the first scroll. The CTA agent called it a decision paralysis risk. One primary action per screen is the standard; we had three.

Copy · High

3,726 words, methodology explained five times

The same methodology explanation appeared across five separate sections: "How it works," "Scoring methodology," "Credibility," "Why it works," and "FAQ." Copy scored 9.5/10 for quality. Length and redundancy were pulling engagement down.

Technical · Medium

Mobile CTA hidden from screen readers

The mobile sticky CTA container had aria-hidden="true" set while remaining focusable. Screen reader users could tab into it but receive no announcement. An accessibility trap, not just a compliance note.

Technical · Medium

Analytics script inflating load time to 73.8 seconds

LCP measured 268ms — the page was fast. But the fullyLoaded metric hit 73.8s because a background analytics script kept the page in a loading state. This doesn't affect user experience directly but distorts performance monitoring.

The fix timeline

We ran four rounds of changes. Each round used the previous audit findings as the brief. The goal was to move Trust above 7.0 and overall above 78.

Round 1
Visual polish — testimonials, credibility, trust
  • Added review carousel with named testimonials and photo placeholders
  • Introduced a credibility bar with trust indicators above pricing
  • Improved visual hierarchy in the hero section
  • Added transparency section with real audit data near the CTA
Round 2
Conversion path optimization
  • Consolidated three above-fold CTAs into one primary action
  • Added pricing anchor near the hero to reduce scroll distance
  • Moved the sample report link to a secondary position
  • Reduced word count by ~600 words through section merges
Round 3
Trust signals, contrast, sticky header, CTA microcopy
  • Added sticky header that appears on scroll (reduces CTA distance)
  • Improved color contrast on calculator labels (2.35:1 → 4.5:1+)
  • Rewrote CTA microcopy to include specific outcome language
  • Added "no subscription" reassurance near pricing
Round 4 — current
Trust emergency: contact, reviews, founder, accessibility
  • Added contact email and support link to the footer
  • Expanded testimonial section to show 4 named reviews with results
  • Added founder section with photo, name, and LinkedIn link
  • Fixed aria-hidden="true" on mobile sticky CTA container
  • Removed analytics script causing 73.8s fullyLoaded time

Before and after

Screenshots captured before Round 1 changes and after Round 4.

Before — March 12, 2026
buyereyes.ai homepage before audit remediation, March 2026
After — Round 4 complete
buyereyes.ai homepage after audit remediation, showing trust improvements

Score comparison

Scores from the March 12 baseline audit vs. the targets we set for Round 4. Re-audit scheduled after all Round 4 changes are live.

DimensionBaseline (March 12)TargetGap
Overall68 / 10078++10
Trust4.5 / 107.0++2.5
Visual5.0 / 106.5++1.5
Copy9.5 / 108.5+ (shorter)reduce length
CTA7.5 / 108.0++0.5
Technical8.1 / 108.5++0.4
Persona avg5.6 / 106.5++0.9

Trust is the bottleneck that limits everything else. A page scoring 9.5 on copy still fails to convert if visitors don't trust who they're buying from. The personas confirmed this — most ended Round 1 in "validation mode," looking for the sample report rather than going to pricing. That behavior shifts when trust signals are present.

Want to see what BuyerEyes finds on your site?

Same audit. 14 agents. 29 sub-scores. Real findings, not generic feedback. Report in 24–48 hours.

View pricing and order a report