Book a call

Meta ads creative testing for ecommerce: a simple system that scales

How to structure tests, what to test first, how much to spend, and when to kill or scale an ad.

By Nathanael Wakefield, founder of Everframe and Elaré 3 min read

On Meta, creative is targeting. The algorithm decides who sees your ad largely from the ad itself, so the brands that win are the ones that test creative fastest and read the results most honestly. Here is the system we run, on client accounts and on our own store, Elaré.

Test in the right order

Most accounts test the wrong things first. Changing a button colour or a headline rarely moves the result. Test from the biggest lever down:

  1. Angle. The reason someone should buy. "Gift that looks expensive" and "jewelry you can shower in" are different angles for the same product, and they find different buyers.
  2. Hook. The first three seconds. Same angle, different opening line or visual.
  3. Format. UGC talking head, product demo, founder video, static, carousel.
  4. Details. Copy, headline, thumbnail, length.

Once you know which angle sells, every new video should be a fresh hook or format on that angle until it stops working.

Structure the account so tests are readable

Keep it simple. One testing campaign and one or two scaling campaigns are enough for most brands under seven figures a month.

  • Testing campaign. New concepts go here, a few ads per ad set, grouped by angle so you can see which idea is working, not just which video.
  • Scaling campaign. Proven ads graduate here with most of the budget behind them.

Do not edit live winners. Duplicate them into scaling. Editing resets learning and can wreck an ad that was working.

How much to spend per test

Give every concept enough spend to show you something real. A useful rule of thumb is to let a test spend at least one to two times your target cost per purchase before you judge it. Less than that and you are reading noise.

Look at leading signals early: thumb-stop rate, hold rate, click-through rate and cost per add to cart. They tell you whether the creative is doing its job before purchases have had time to add up.

When to kill, when to scale

Kill an ad when it has spent past your threshold with weak leading signals and no purchases. Do not keep feeding a loser because the video "should" work.

Scale an ad when it beats your target cost per purchase over several days, not one lucky afternoon. Raise budgets in steps, and move the winner into the scaling campaign rather than pushing a test ad set too hard.

Iterate on anything in between. An ad with a great hook rate and a weak click-through often has the right opening and the wrong offer or ending. Recut it rather than killing it.

Check the platform against your store

Meta's reported purchases are modelled and include view-through conversions. Your Shopify orders are what actually happened. The two will not match, and that is normal. What matters is that you compare them consistently, every week, and make budget calls on the trend in your own store data, not on a single day of platform ROAS.

Keep the winners fed

Every winning ad fatigues. Frequency climbs, cost per purchase creeps up, and performance fades. The answer is volume: a steady flow of new hooks and formats on your proven angles, so there is always something ready to replace the ad that is tiring.

This is where most brands fall over. The media buyer finds a winner, nobody makes the next video in time, and the account stalls. That is the reason Everframe runs content and ads as one team.

What good looks like

On accounts we have run, the peak has been a 17.11x ROAS on a single ad set, and $225K at 4.86x in 15 days with $75K net profit. Those are peaks, not promises. The repeatable part is the system: test angles first, judge on enough spend, scale winners carefully and never stop feeding the account.

If you would rather have someone run this for you, see how we approach Meta ads, or read the Effortless Shop case study, a $631,845 month at a 10.11% store conversion rate while we ran TikTok and Meta end to end.

Work with Everframe

Want this run for you?

We make the content and run the Meta and TikTok ads for ecom brands, or train your team to do it in-house. Plans from $3,000 a month. Every content package is guaranteed 100K views, or your money back.

Keep reading

Book a 15-minute call