Landing page friction teardown
Use when a landing page underperforms and you need a structured list of friction points rather than opinions.
Fill in before running
Replace each placeholder with your own detail. The more specific you are, the less the model invents.
- {{PAGE_COPY_OR_SCREENSHOT_DESCRIPTION}}
- {{PRIMARY_ACTION}}
- {{AUDIENCE}}
- {{TRAFFIC_SOURCE}}
Getting a better result
- Paste the ad or email copy that sent traffic to the page, otherwise message match cannot be judged.
- Give the price and commitment level, since most friction findings depend on what the visitor is being asked to risk.
- Run the same prompt against a competitor page and compare the severity tables.
Questions about this prompt
When do I run a teardown rather than commissioning a redesign?
When the page underperforms and the internal debate has turned into competing opinions. The seven point checklist runs in a fixed order starting with message match, so every finding arrives tied to a named item, a severity and a confidence rating. If your doubt is only about the first screen, the above the fold check is narrower and faster.
What do I need in front of me before running it?
The page copy or a description of the screenshot, the primary action, the audience and the traffic source. Paste the ad or email copy that sent the traffic as well, because message match cannot be judged without it. Include the price and the commitment level, since most friction findings turn on what the visitor is being asked to risk.
What comes back, and which part is worth acting on?
A table of checklist item, what was observed, severity, suggested change and confidence, sorted highest severity first, then the three changes to make first and a separate note of which items passed. The rows marked cannot assess without data are the useful ones, because they are your list of what to go and measure.
What is the mistake that costs me here?
Describing the page instead of pasting it. A summary in your own words already contains your judgement about what matters, and the reading order and visual hierarchy findings depend on the sequence you supply. Running it on your own page without running it on a competitor page leaves you no sense of which severities are normal.