The takeaway: Cost per app install usually falls when you find the right opening and message for your audience, and that only happens through continuous testing. 10 new videos a week give the algorithm enough opportunities to find a winner.

Why are apps different?

  • Users usually do not pay when installing, so the decision is very quick and emotional.
  • Competition for the same audience is very high on TikTok and Snapchat.
  • Ads fatigue quickly because app audiences see so many ads.

What do we test each week?

  1. The opening: A problem, curiosity, or a result.
  2. The screen demonstration: Someone holding a phone and showing the app in action.
  3. The first benefit: What the user gets in the first minute.
  4. The closing: "Download it for free," "Get a discount on your first order."
Person holding phone showing clothes and products with chat bubbles shopping bags and payment cards for weekly testing
Weekly testing of products messages shopping bags and payment methods. · AI illustration

A plan for 10 videos a week

  • 5 new openings with the current best-performing message.
  • 3 different characters (age, gender, dialect).
  • 2 completely new ideas.

At the end of the week: stop the weakest performers, double the budget for the best, and add 10 new ones.

What do you measure?

  • The percentage watching the first 3 seconds.
  • Click-through rate to the app store page.
  • Cost per install (CPI) and the cost of the first in-app transaction.

We do not promise a specific percentage reduction; results depend on the app, market, and budget. But faster testing means finding the winning ad sooner. See the testing matrix for details.

Which openings work for app ads?

The strongest opening in an app ad shows the app solving a problem within two seconds, not the app's logo. We usually test four kinds:

  1. Screen first: a screen recording of the app completing the task.
  2. Everyday problem: a situation the user knows, then the fix inside the app.
  3. Comparison: the old way versus the app.
  4. Offer: a first-order discount or free trial, stated clearly.

What do we need from you for app ads?

  • Real screen recordings of the app, or a test account we can record from.
  • The app's App Store and Google Play links.
  • The feature you want each campaign to focus on.
  • The current offer and its terms.

We never show screens or features the app does not have, because that drives uninstalls and breaks platform policies.

Shop staff handing shopping bags and order with checklist while customer shows products on phone for app ad
Document your products and orders by phone to prepare app ad materials successfully · AI illustration

How should you run a testing week for app ads?

Split the week into launch, read and decide. Launch new versions at the start of the week with equal budgets, read results once each version has enough data, then pause the weak ones and move budget to the strongest. Testing becomes a weekly habit instead of a random experiment.

Start of the week
Step
Launch new versions with equal budgets
Mid-week
Step
First review, without pausing too early
End of the week
Step
Pause the weak versions and move budget to the winner
Before the next week
Step
Order new hooks built on the winner

Which mistakes raise the cost per install?

Most mistakes come from an ad that promises something the user cannot find in the app, or from a hook that explains too much before showing the benefit. Both raise costs, because the viewer either does not finish the video or installs the app and quickly leaves.

  • The app logo in the first second instead of the benefit.
  • Screens that do not exist in the current version of the app.
  • One ad for every audience instead of one version per use case.
  • Pausing versions before they have gathered enough data.

Why should an app ad match its store page?

An app ad brings the click, but the install decision happens on the app's App Store or Google Play page within seconds. If users find the same benefit and the same screen that drew them in the video, they install. If they find a different message, they hesitate and leave, so a click you paid for is wasted and your cost per app install rises.

The benefit in the first two seconds
What should match on the store page
The same benefit in the first store screenshots and the first line of the description
The screen shown in the video
What should match on the store page
The same screen among the store screenshots
An offer such as a first-order discount
What should match on the store page
The same offer on the same terms
The ad's language
What should match on the store page
A description in the same language, Arabic or English

Frequently asked questions

Can the app appear in the video without filming? Yes. We place the real screen recordings inside phone scenes, with a digital character explaining how to use it.

How many versions should I test each week? Start with several openings for the same feature, keep the best, and add new openings every week.

Which number matters most in app campaigns? Cost per install, then the share of users who complete sign-up or their first action after installing.

Do the same ads work for iOS and Android? Usually yes, as long as the screens shown match the system the ad runs on.

Which language should app ads in the Gulf use? It depends on your audience: Gulf dialect versions for local users and English versions for residents. Test both.

Do app videos work on TikTok, Snapchat and YouTube Shorts at once? Yes. We deliver every video vertical in 9:16 at 1080x1920, ready for Meta, TikTok, Snapchat and YouTube Shorts ads.

Can I try a video for my app before paying? Yes. We make a free sample video before payment, and the first batch then arrives within 48 to 72 hours of script approval.

Part of our approach to Video Ads for Salla, Zid, and Shopify Stores.