Input-driven hero matches the category pattern (v0, Lovable, Tome all do the same — paste something, get the output). For an AI generator, this is correct: the product starts before signup.
The example prompt inside the box teaches the use case without a separate explainer section: "Convert this TED talk into presentation slides · Ratio [16:9] · [10-12 pages]." One line, three parameters, exactly what the model needs.
Visual register is premium — serif headline, muted beige ground, floating sample card. Reads like a tool a design-conscious PM would trust, not a hack-week project.
Top 3 conversion killers
Killer 1
The H1 promises nothing
"What will you present today" is a greeting, not a pitch. There is no category word (presentation, deck, slides), no outcome (ready-to-present, export to PPTX, branded), no time (90 sec, 2 min), and no audience (founder, sales, PM). A first-time visitor has to squint at the input-box placeholder to reverse-engineer what PageOn even is. Every cold visitor who doesn't already know the product bounces in under 3 seconds.
Killer 2
The embedded card in the H1 breaks comprehension
Interrupting an H1 with a visual element forces the eye to re-enter the sentence on the other side. It also implies the card is the point. On desktop it's parseable; on mobile it breaks the sentence across lines and looks like a bug. A strong hero earns a quick visual accent — it doesn't build the accent into the one sentence that has to sell.
Killer 3
The placeholder URL is a rick-roll — trust cost is real
About 20% of visitors will recognize it and smirk. The other 80% will read it as either a broken URL or an unprofessional choice. For a product that costs money and targets people putting slides in front of colleagues/clients, the first piece of micro-copy on the page should not be an in-joke. The confidence cost is bigger than the engagement win.
Hero rewrite
Before
H1: What will you [card] present today
Sub: (none)
Placeholder: youtube.com/watch?v=dQw4... · "Convert this TED talk into presentation slides"
After
H1: Turn any link, PDF, or brief into a ready-to-present deck in 90 seconds.
Sub: Paste a YouTube talk, a doc, or a bullet outline. PageOn writes the narrative, designs the slides, and exports a branded deck you can open in Keynote or PowerPoint. No templates, no copy-paste.
Placeholder: Paste a link, drop a PDF, or describe the deck you want…
Example chips: Notion doc → deck · YouTube talk → deck · Bullet outline → deck
CTA (secondary): Watch a 60-second demo
Why the rewrite works: "90 seconds" is a concrete, testable promise. "Ready-to-present" anchors the outcome. "Keynote or PowerPoint" names the actual exit format every PM and founder needs. Example chips replace the single rick-roll placeholder with three real use-cases.
Navigation fix
Before
Top nav: Features · Pricing · Blog · Book a Call · English · Log in
After
Top nav: Features · Pricing · Blog · Log in
Move "Book a Call" to footer under "Enterprise / team plans." In self-serve SaaS, "Book a Call" in the top nav raises friction and signals "this product isn't ready to stand on its own." Guide the 95% who are about to try it, not the 5% who want a call.
A/B test to run next
Control: "What will you present today" · rick-roll placeholder URL
Variant: "Turn any link, PDF, or brief into a ready-to-present deck in 90 seconds." · three example chips (Notion / YouTube / outline)
Primary metric: Input-box submissions from the hero (the only conversion event on this page).
Secondary: Signup rate on the post-generation gate.
Sample size: ~1,500 hero visits per arm for 95% CI on a +25% relative lift.
If you can't swap both at once, ship the placeholder change first — it's a one-line edit and the trust cost of the rick-roll is the fastest-compounding loss on the page.
This teardown is free.
The full $79 version covers hero + sub + CTA + feature section + one A/B test, delivered as a PDF in 48h. Reply to benjamin@pagerewrite.com with your URL.