Back to blog
Performance Marketing

Meta Ad Creative Performance: Read Each Asset's Results

Meta Ad Creative Performance: Read Each Asset's Results

Meta ad creative performance can now be read asset by asset. A new breakdown in Ads Manager splits results for each image or video delivered as single media and for each text variation, so advertisers who load several assets into one ad can finally see which one earned the clicks and conversions instead of guessing (Social Media Today, as of October 27, 2026).

What changed in Meta Ads Manager reporting?

Until now, an ad with three images and five copy options reported as a single line. Meta's delivery system knew which combinations worked, but the advertiser did not. The new breakdown opens that box.

BreakdownWhat you seeExample
MediaEach image or video served as single mediaProduct-only shot vs. in-use video
TextEach primary text or headline variationDiscount-led copy vs. review-quote copy

Open the Breakdown menu in the Ads Manager table and pick the creative option. Impressions, clicks, conversions and spend then appear per asset.

Which ads are left out of the breakdown?

Not every ad qualifies. Meta lists two exclusions:

  1. Ads opted into multi-post formats, such as "Show as collection" or "Show as carousel", are not split by asset.
  2. Some custom conversions, for example "Donate" and "Subscribe", are not supported.

Many accounts leave automatic carousel conversion switched on by default. If the goal of a test ad is to compare assets, turn that option off for the test, or the breakdown will stay empty.

Why it pairs with creative-level targeting

One week earlier Meta added audience settings for each creative and retired detailed targeting. With age and gender now attachable to individual assets, and results now readable per asset, advertisers can test a direct claim: this asset works for this audience. We covered the setup in Meta Creative Targeting: Match Assets to Audiences.

Takeaways for marketers

  • Write a replacement rule before you look. For example, retire the asset with the worst cost per conversion inside each ad once it has enough volume. A written rule stops debates driven by personal taste.
  • Judge after learning, not during it. Adding or removing assets can push an ad back into learning. Use the threshold in Meta Ads Learning: Why 50 Conversions a Week Matters to time decisions.
  • Make copy testing routine. In many accounts the gap between text variations is wider than the gap between images. Now that text results are visible, a monthly hook-line test is cheap to run.
  • Add one creative table to every report. Ad-level reporting rarely helps the production team. An asset-level table turns the next shoot brief into a numbers conversation.

Our view at BestPartner: decide what to cut before you add more assets. The faster weak assets leave, the longer the remaining ad can stay out of learning on the same budget.

Source: Social Media Today

Frequently Asked Questions

Where do I find creative-level results in Meta Ads Manager?

Open the Breakdown menu in the Ads Manager table and choose the creative option. Results then appear for each image, video and text variation.

Does the breakdown work for carousel ads?

Ads opted into multi-post formats such as Show as collection or Show as carousel are not split by asset. Turn that option off for test ads.

Are all conversion events supported?

No. Some custom conversions, including Donate and Subscribe, are not covered as of October 2026.

Should I swap weak assets immediately?

Wait until the ad has exited learning. Changing assets can restart learning and blur the comparison.

Where does your own site stand?

To apply what you just read to your own site, start with a free audit of where things are now.

A strategist replies within one business day.

Read next

Measuring AI Usability: Why Token Counts Mislead
AI Marketing모비인사이드

Measuring AI Usability: Why Token Counts Mislead

Measuring AI usability means tracking how work changes, not how many tokens a team burns. Four things that inflate usage and a four-step way to measure.