AI Prompt for Product Title A/B Variants

Product title A/B test ideas: Stop guessing which title wins. Comes with a usage guide and three practical tips. Takes under a minute to customize.

The prompt

You are a conversion-rate optimization specialist who designs clean A/B tests for product titles.

Context: My current title is [CURRENT_TITLE] for [PRODUCT] at [STORE_NAME]. It converts at [CURRENT_CONVERSION] and I want to beat it. My buyers are [TARGET_CUSTOMER].

Task: Write [NUMBER] alternative titles, each testing a DIFFERENT angle: one keyword-led, one benefit-led, one social-proof-led ("loved by 10,000+"), one specificity-led (exact numbers), one curiosity-led, and one urgency-free scarcity cue. No two variants may use the same angle.

Format: A numbered list. For each variant: the title, the angle it tests (in bold), and a one-line hypothesis ("If X, then Y because Z").

Constraints:
- Every variant under 120 characters.
- Keep brand and core keyword constant across variants so the test is fair.
- Only one variable changes per variant - the angle.
- No false claims; social proof numbers must be ones I can verify.
- Write a 3-sentence test plan at the end: traffic split, duration, and success metric.

Variables to fill in

Replace each [variable] in the prompt with your own details:

[CURRENT_TITLE] [PRODUCT] [STORE_NAME] [CURRENT_CONVERSION] [TARGET_CUSTOMER] [NUMBER]

Example input

CURRENT_TITLE: Insulated Water Bottle 32oz | PRODUCT: insulated bottle | STORE_NAME: TrailPeak | CURRENT_CONVERSION: 2.1% | TARGET_CUSTOMER: hikers | NUMBER: 6

How to use this prompt

Run one variant against the control at a time with 50/50 traffic for at least 7 days or 200 conversions per side, whichever comes later. Change only the title - touching images or price at the same time ruins the test.

Tips for the best results

Test angles, not adjectives. 'Benefit-led vs keyword-led' teaches you something; 'great vs amazing' teaches you nothing.
Let losing variants die fast - move 80% of traffic to the winner after statistical significance.
Keep a log of every test and result; after 10 tests, a house style guide emerges that no competitor can copy.