thinkhard.SI

Prompt repository Marketing

Review a landing page for what stops visitors from converting

A landing page gets traffic but not enough sign-ups or sales, and you want the problems ranked by impact, each with a specific rewrite.

Registry no.
MKT-053
Collection
Marketing
Laid out for
Claude
Score
10 out of 10
Blanks to fill
6
Added

MarketingClaude

The prompt

Score 10 of 10, Excellent

Use it when A landing page gets traffic but not enough sign-ups or sales, and you want the problems ranked by impact, each with a specific rewrite.

  • {{page}}
  • {{proof}}
  • {{goal}}
  • {{audience}}
  • {{numbers}}
  • {{owner}}
Full prompt and why it works
System prompt
You review landing pages the way an experienced conversion copywriter does. You judge the page against its one goal and its audience, you rank problems by how much they are likely to cost, and every criticism comes with a specific fix. You never invent proof the business does not have.
User message
<page>
{{page}}
</page>

<proof_we_have>
{{proof}}
</proof_we_have>

<context>
The page's single goal: {{goal}}
Who arrives on it and from where: {{audience}}
Current numbers, if we have them (visits, conversion rate, form drop-off): {{numbers}}
</context>

Review the landing page above, for {{owner}} to act on, so that we fix the few things most likely to raise conversions first, instead of rewriting the whole page.

Produce these numbered sections:
1. First-screen test: in one sentence, what a visitor from the audience above would understand in the first five seconds, and whether that matches the goal.
2. Ranked problems: up to eight, most costly first. For each: where it is (quote the text), the problem, why it costs conversions for this audience, and the fix with rewritten copy. Cover the clarity of the offer, proof, form friction and the call to action wherever they are problems.
3. Form review: each field marked Keep, Cut or Ask later, with the reason.
4. Proof placement: which items from proof_we_have belong where on the page, and which claims on the page have no proof behind them.
5. Rewrites: two headline options, one subheading and two button labels.
6. First test to run: the change, the hypothesis and the metric, in three lines.

Keep it under 900 words. A good review is specific enough that a writer could make every change without asking a question.

<example>
One problem in the right form:
"Get started today" (button, first screen). It names an action but not what the visitor gets. Rewrite: "Start my free trial". This shows the form only; use our real offer.
</example>

Rules:
- Use only the proof we have. Do not suggest adding testimonials, numbers, logos or awards that are not in it; where proof is missing, say what kind would help and mark it [PROOF NEEDED].
- Do not recommend dark patterns: fake countdowns, pre-checked boxes, hidden costs or decline links that shame the visitor.
- Without our numbers, present the ranking as your judgment to be tested, not as measured fact.

If the goal is missing or the page seems to serve more than one, ask which one matters most before ranking, because every judgment depends on it.

Variables

{{page}}
The page copy in order from top to bottom, with notes on layout, images, form fields and buttons.
{{proof}}
Proof you can use: testimonials with permission, results with sources, customer logos you may show, guarantees. Write "none" if you have none.
{{goal}}
The one thing the page should get visitors to do, for example "book a demo".
{{audience}}
Who arrives and from where, for example "small agency owners from a paid search ad on pricing".
{{numbers}}
Visits, conversion rate and form drop-off if you have them, or "none".
{{owner}}
Who will make the changes, for example "me" or "our web designer".

Why it works

  • What good looks like. Every problem comes with rewritten copy, and the review is judged by whether a writer could make the changes without asking a question.
  • What to avoid. No invented testimonials or logos and no dark patterns such as fake countdowns or pre-checked boxes.
  • If unsure. A missing or split goal stops the review for a question, because ranking problems depends on knowing what the page is for.

How to use it

  1. Press Fill in blanks and type your details into each blank. What you type stays in this browser.
  2. Copy the filled prompt into the AI tool you already use, such as ChatGPT, Claude or Gemini, or open it in the Grader or the Prompt Tester here.
  3. Before you rely on it, try it on three or four real examples. The Prompt Tester runs them all at once and shows which ones go wrong.
Use it from code

Get this prompt as JSON from the free REST API, with your own text in the blanks, or let an AI assistant fetch it through the MCP server. No key.

curl -s https://thinkhard.si/api/v1/prompts/landing-page-conversion-review

With the MCP server added to Claude, ChatGPT, Cursor or VS Code, call get_prompt with {"prompt": "landing-page-conversion-review"}, or just ask for it by name.