The 3 leaks costing them the most
1 The hero needs a decoder ring
Why it hurts: 'Think fast, build faster / Brainstorm in chat, build in Cowork' assumes you know what Cowork is (#20). A first-timer can't tell what they'll actually do here.
Fix: Add a plain line: 'Chat with Claude to write, code, analyze, and get real work done.'
2 It tries to be everything
Why it hurts: Claude, Claude Code, Cowork, Security, Chrome, Slack, M365, Skills (#11) β the product menu sprawls before the visitor has done one thing.
Fix: Lead with the single core use, then reveal the platform.
3 Free-led, login-first, barely shown
Why it hurts: A free plan fronts a login-first funnel (#1, #8) and only 3 images appear (#10). For a visual, interactive product, that's almost nothing.
Fix: Show Claude actually doing something impressive before asking anyone to sign in.
All 31 principles, scored
A free plan ($0) fronts the funnel β the rubric's weak-monetization flag.
Fix: Lead paid plans with a trial framing.
Only 5 distinct colors detected β a clean palette.
No concrete numbers in the headings.
Fix: Quantify a real productivity gain.
Large, standard footer.
Fix: Add a hook.
Has an og:image and a clear OG description.
Fix: Make the OG title more than just 'Claude'.
Products, features, models, solutions stack many ideas per screen.
Fix: Cut the homepage to one idea.
'Think fast, build faster' is plain and readable.
Login-first ('Continue with Google') freemium funnel.
Fix: Show value before the login wall.
'Think fast, build faster / Brainstorm in chat, build in Cowork' is somewhat ownable.
Fix: Add a line only Claude could write.
Video exists but only 3 images β the product is barely shown.
Fix: Show Claude doing impressive work up top.
Claude, Code, Cowork, Security, Chrome, Slack, M365, Skills sprawls wide.
Fix: Anchor on one wedge use case.
Free/Pro/Max/Team/Enterprise exceeds the popcorn limit.
Fix: Show three tiers; hide enterprise.
Claude IS the AI wave.
'do your best work', 'think fast' is user-ish, if abstract.
Fix: Use the words people say about what Claude helps them finish.
No founder or human presence.
Fix: Add a human voice.
'Explore plans' / Pricing exists, though not a clear top-nav item in the capture.
Fix: Add a plain 'Pricing' nav link.
'Think fast, build faster' is reasonably recallable.
Fix: None needed.
Mild aspirational pull.
Fix: Add a sharper hook.
Cowork and Skills are genuinely new surfaces.
Fix: Spotlight the never-seen capability.
The hero conveys 'AI for work' but leans on knowing what Cowork is.
Fix: Add a plain explainer line.
Jumps to login before naming a felt pain.
Fix: Open with the work-overwhelm pain.
Multiple login options and CTAs compete.
Fix: Pick one primary action.
'Claude' is a distinct, human, memorable name.
Sells 'do your best work' β desire, if abstract.
Fix: Make the win concrete.
Free plan lets you try it in chat.
Fix: Embed a live prompt-to-output demo.
Only one weak word detected.
Subscription model.
Fix: Not realistically one-time.
'Continue with Google' is a login, not an outcome; 'Try Claude' is better.
Fix: Lead with 'Try Claude'.
Testimonial markup exists but no visible quotes or avatars.
Fix: Add named customer stories.
'Next generation AI assistant built by Anthropic' is under ten words.
Fix: Put a crisp line in the hero.
The Max tier ($200) signals premium, though free entry anchors low.
Fix: Make the premium value explicit.
How would your page score?
Same 31 principles. Same brutal honesty. Free.
Grade My Page