APromptAura
0

Ad variant generator with a real testing hypothesis

auto-checked, 4 days old
AauraNovice

The prompt

0 copies
Generate number ad variants for platform promoting product_offer to target_audience, but structure this as an actual test, not just a pile of random copy. Core message/offer to test: core_message Budget/format constraints: format_constraints Do this: 1. Before writing variants, define 3-4 distinct hypotheses about what might drive clicks/conversions here (e.g. 'leading with price beats leading with outcome,' 'urgency beats social proof for this audience,' 'a question headline beats a statement headline'). Name each hypothesis explicitly. 2. Write 2 variants per hypothesis, differing ONLY on the variable being tested (so results are actually attributable), keeping everything else as similar as possible. 3. For each variant, state which hypothesis it tests and what result would confirm or kill that hypothesis. 4. Flag if any variant makes a claim that would need substantiation/compliance review (health claims, guarantees, superlatives) before running. 5. Recommend a minimum sample size or budget split logic for a fair test, in plain terms (not exact statistics, just sound testing logic).

Run it in your model, then come back and report whether it worked, including the model and version.

Where it has been run

People and our automated sweep are counted separately. Only the People column is someone vouching for this prompt.

ModelVersionPeopleBrokeAutoLast
ClaudeHaiku 4.5001/14 days ago

Example output

4 hypotheses defined: (1) outcome-led beats feature-led headline, (2) urgency beats social proof, (3) question headline beats statement headline, (4) specific number beats vague benefit. 8 variants generated, 2 per hypothesis, each pair identical in format/CTA/image direction, differing only on the tested variable. Example pair for hypothesis 1: Variant A 'Cut your invoice time to 5 minutes' (outcome), Variant B 'Automated invoicing for small teams' (feature), same CTA button, same visual direction, so any performance gap is attributable to the headline framing alone. Compliance flag: variant testing hypothesis 4 uses the number '90% faster,' flagged as needing a substantiation source before running. Testing logic note: recommend running all pairs simultaneously with equal budget split rather than sequentially, since day-of-week and audience fatigue effects would otherwise confound results.

0 comments

Sign in to comment or report what this prompt did for you.

Sign in

No comments yet. Run the prompt and report what you got.