Ad Creative Testing Checklist: Before, During and After Each Test

• Start from a problem, not a whim. What's weakest in your current best ad: the hook, the offer, the proof, the format?
• Prioritize by impact. Hooks and offers affect everyone who sees the ad. Small details, like button wording, usually matter less.
• Test big differences first. Different angles, formats or offers teach you more than small wording tweaks, especially early on.
• Check the basics. Tracking works, the landing page matches the ad, the offer is clear. Otherwise your test measures those problems instead.

A hypothesis turns a hunch into something you can learn from. A simple format:

Because [observation], we believe [change] will [expected effect] for [audience]. We'll judge it on [metric] after [spend or time].

For example: “Because our current ad opens on the product rather than the problem, we believe a problem-first opening will raise CTR with cold audiences. We'll judge it on CTR and cost per purchase after equal spend on both versions.”

• Change one variable. If the hook, visual and offer all change, you won't know which one mattered.
• Keep everything else equal: audience, placements, budget, schedule and landing page.
• Run versions at the same time, so day-of-week and seasonal effects hit both equally.
• Use your platform's testing tools where they fit, such as Meta's A/B test feature, to split audiences cleanly.
• Name ads clearly so you can tell versions apart later, for example “Mask_HookPain_v1”.

• Primary metric: usually cost per purchase or cost per lead, with CTR and early video metrics as supporting evidence.
• Minimum data: how many clicks or conversions, or how much spend, each version gets before you judge it. Small numbers produce confident-looking results that don't hold up.
• Time limit: when you'll stop if there's no clear difference.
• What counts as a win: a meaningful difference, not a tiny one that could be noise.

• Don't edit ads mid-test. Significant edits can reset delivery and muddy the results.
• Don't declare a winner after the first day. Early results swing a lot.
• Watch for delivery imbalances, where one version gets most of the spend, and note it when reading the results.
• Read the comments for objections and questions. It's free qualitative feedback.

1. 1Compare versions on your primary metric first, then use supporting metrics to understand why.
2. 2Ask whether the difference is big enough to trust, given how much data you have.
3. 3Check the whole funnel: did the winning hook bring clicks that converted, or just more clicks?
4. 4Decide: scale the winner, iterate on it, or go back with a new hypothesis.

Keep a simple test log. For each test, note the date, hypothesis, what changed, the results, and a one-sentence conclusion. Over time the log becomes one of your most valuable creative assets: a record of what your audience actually responds to.

Include the losing tests. “Discount-led hooks didn't beat problem-led hooks for this product” is a lesson you won't have to pay for twice.

A structured review is most useful at the start of this process, when you're deciding what to test. The ad creative analyzer assesses ten parts of an ad, flags the weakest, and writes a testing plan with a hypothesis for each test that you can drop straight into your log. To review many ads at once, see the ad creative audit process.