Fill in these first
- Platform: [search ads, social ads, display]
- Current ad copy: [paste the headline, body and call to action]
- Current results: [impressions, click-through rate, conversion rate, cost per result]
- Audience: [who sees the ad]
- Known objections or feedback: [comments, sales notes, reviews]
- Weekly budget: [budget in INR or USD]
Task
Write five test hypotheses for this ad. Each test changes exactly one variable so the result can be read clearly.
Variables to choose from
Headline angle, offer framing, proof type, call to action wording, specificity of numbers, pain versus gain framing, audience call-out in the first line.
Output format
| Rank | Hypothesis | Variable | Control | Variant | Primary metric | Win condition |
Write each hypothesis as: "If we change [variable] to [variant], then [metric] will improve because [reason based on the inputs]."
After the table:
- Test order: which test to run first and why, in two sentences.
- Sample size note: the rough number of clicks or conversions needed per version before reading the result, given the budget.
- Stop rule: when to end a test early.
Quality rules
- One variable per test. If a variant changes two things, split it.
- Reasons come from the inputs, not general advice.
- Variant copy obeys the platform's character limits.
- Rank by expected impact multiplied by ease of running.
- No variant makes a claim the control could not also make truthfully.






