End cards are the most under-tested asset you own

This piece is now part of a longer guide: Playable ads: the complete guide.

Everyone tests hooks. Almost nobody tests end cards, which is strange, because the end card is the frame that is on screen when the decision actually happens.

Where it sits in the unit

On AppLovin the sequence is video, then one end card, then the destination. The end card is not decoration and it is not a logo slide. It is the handover. The video builds the reason, the end card converts it into an action, and the page has to keep the promise both of them made.

Why it goes untested

Because it is produced last, usually by whoever is closest to the timeline, and because it feels like packaging rather than creative. In most accounts a single end card is reused across an entire library, which means one asset is carrying the conversion moment for every concept you run.

What is worth varying

• The claim. Repeat the video's specific promise, or state the offer. These perform differently and neither is universally right.
• The proof. A rating, a count, a before and after.
• The action. What the button says, and whether it names the next step or the outcome.
• The register. Product-forward versus lifestyle. A bake-off between those two is the single most informative end-card test we run.

It got cheaper in June 2026

AppLovin added built-in end-card creation to Ads Manager, generating a card from a png, jpg or gif with an automatic CTA, plus a Media Library for bulk asset uploads. The production excuse for reusing one card across a library is largely gone.

What to run first

Hold the video constant. Run three end cards against it. You will learn more about what your audience responds to in a week than another round of hook testing will tell you, because you are testing the moment of decision rather than the moment of attention.

Questions people ask

Does the end card need to match the video visually? It needs to match the promise. Visual continuity helps, continuity of claim matters more.

Can we use one card for everything? You can, and most brands do. It is also why end-card performance is invisible in most accounts: with no variation there is nothing to compare.