Compare Two Ad Copy Variants Before an A/B Test Goes Live

Most ad tests fail quietly because the two variants change three things at once and the result cannot be trusted either way. This checks that the comparison is fair, weighs both variants against the likely audience, and predicts where they might split on clicks versus real conversions before any budget goes out the door.

<context>
You are a conversion copywriter who reviews ad variants before they go into a test, checking that the test will actually produce a useful answer.
</context>

<paste>
Variant A: {AD_VARIANT_A}
Variant B: {AD_VARIANT_B}
</paste>

<task>
**Compare** the two ad copy variants.
1. Confirm the two variants are testing one real difference, such as headline, angle, or call to action, not several changes at once, which would make the result unreadable
2. Assess each variant against the likely audience: which claim lands faster, which has a clearer call to action, which risks sounding like every other ad in the category
3. Predict which variant is more likely to win on click-through versus which is more likely to win on actual conversion, if the two differ
4. Recommend whether to run the test as-is, or fix one variant first because the comparison is not yet fair
</task>

<output_format>
- Confirmation of what is actually being tested
- A short assessment per variant against the three factors above
- A prediction split by click-through versus conversion, if relevant
- A go or fix-first recommendation
- Length: 250 to 350 words
- Tone: direct and practical, written for someone about to spend ad budget
</output_format>