How to Test Paid Social Creative (and What Budget You Actually Need)

Why most creative tests are inconclusive

Three problems show up in almost every paid social creative test:

• The audience changes between variants (Meta optimises differently each time)
• The budget is too small to clear statistical noise
• The "winning" creative is really the one Meta chose to show to the easiest converters

Fix those, and creative testing actually tells you something.

A simple, defensible test structure

1. Hold the audience constant. One Custom Audience per test. No interest-stacking, no Advantage+ swapping the audience under you.
2. 3–5 creative variants per test. Same offer, same landing page. Vary one dimension (hook, format, message).
3. Use a dedicated test campaign. Don't pollute your scaling campaigns.
4. Pick a primary metric per stage. Hook rate / thumbstop for top-of-funnel; CTR for mid-funnel; CPA for bottom-of-funnel.

What budget do you actually need?

Rules of thumb:

• Hook tests: ~£50–£100 per variant to read 3-second view rate confidently
• CTR / engagement tests: ~£150–£300 per variant
• Conversion tests: at least 30–50 conversions per variant, back-calculated from your CPA

If your CPA is £40 and you want 30 conversions per variant, you need ~£1,200 per variant just to read the test. That's why most "creative tests" don't actually conclude.

Why audience choice decides the winner

Show two creatives to slightly different audiences and the "winner" is the audience, not the creative. The fix is a fixed, well-defined Custom Audience. Ideally one that matches the customer the creative is for. A creative built for recent movers should be tested against a recent-mover audience, not Meta's broadest pool.

How Outra makes creative tests fair

Creative testing is only meaningful when the audience is stable and intentional. Outra resolves your customers to a household identifier and pushes named, household-level Custom Audiences into Meta. Just moved, Owner-Occupied Homes, High Purchasing Power, Bigger Families. You can finally test creative for a specific household type against that exact household type. See also: how to enrich Klaviyo profiles with the same household attributes.

The result: tests that conclude, learnings that compound, and creative spend aligned with the audiences your brand actually wants. Audience targeting is the lever paid social removed. Outra puts it back so the creative your team invests in lands on the right households.