ADS & LANDING PAGES

CTA rewriting with a friction check

Use when buttons and calls to action are generic and clicks are dropping at the final step.

cta-rewriting-with-friction-check.md

Download .md

You are a conversion copywriter working only on calls to action.

Page or ad: {{PAGE_OR_AD}}
Current CTA labels and where each sits: {{CURRENT_CTAS}}
What actually happens after the click: {{WHAT_HAPPENS_NEXT}}
Cost or commitment at that point: {{COMMITMENT_LEVEL}}
Audience: {{AUDIENCE}}

Step 1: For each existing CTA, judge it against four tests and mark pass or fail with one line of reasoning: does it name the outcome, does it match what actually happens next, is the commitment level honest, is it in the visitor's language.
Step 2: Rewrite each CTA with three options - one outcome-led, one low-commitment, one specific and literal. Max 5 words each.
Step 3: For each rewrite, write the microcopy that sits beneath it (max 12 words) removing the fear that stops the click.
Step 4: Flag any CTA where the copy promises less friction than reality. This is the most damaging problem, so list it first.

Output as a table: Location, Current, Option A, Option B, Option C, Microcopy, Honest? (yes / no).

Rules:
- Never write a CTA that implies no card is needed if a card is needed.
- Do not use "Submit", "Click here", "Learn more" or "Get started" unless you justify it in one line.
- No em dashes.

Fill in before running

Replace each placeholder with your own detail. The more specific you are, the less the model invents.

{{PAGE_OR_AD}}
{{CURRENT_CTAS}}
{{WHAT_HAPPENS_NEXT}}
{{COMMITMENT_LEVEL}}
{{AUDIENCE}}

Getting a better result

Describe what really happens after the click, including any card or call requirement.
Match the CTA commitment to page position - the top of the page should ask for less.
Test the microcopy separately from the button label; it often moves more than the label.

Questions about this prompt

When is this better than just writing punchier button copy?

When clicks fall away at the final step. Rewriting labels alone gets you a livelier verb and the same drop off, because the problem is usually the distance between what the button implies and what happens next. This tests each CTA against that reality and writes the microcopy underneath, which often moves more than the label does.

What do I need in front of me?

Every current CTA with its position on the page, and an honest account of {{WHAT_HAPPENS_NEXT}}. Say whether a card is required, whether a sales call follows, how long the trial runs. The honesty check is the point of the prompt, and it cannot flag a CTA promising less friction than reality if you described the friction as none.

What does it return?

A table per location with three rewrites, outcome-led, low-commitment and literal, plus microcopy and an Honest yes or no column, with any dishonest CTA listed first. That Honest column is worth more than the labels. A button implying no card when a card is needed converts nicely and generates refunds and complaints later.

What goes wrong?

Rolling one winning label across every button on the page. The top of the page should ask for less than the bottom, and the same high-commitment label repeated reads as pressure. Test the microcopy separately from the label as well, or a winning variant tells you nothing about which of the two actually did the work.