App Promo Video: What to Show, What to Cut, and Where It Runs

Most app promo videos fail for a dull reason: someone made one polished export and pasted it into the store listing, the ad account, the landing page, and YouTube. Four surfaces, four different viewers, one file. The store version feels like an ad. The ad version feels like a manual. Nobody installs.

The fix is not more production budget. It is deciding what earns screen time before you record, then cutting that same footage three different ways.

Start with one workflow, not a feature list

A good app promo video shows a person getting one thing done inside your app. Not a tour. Not a menu. One complete job, from intent to payoff.

Apple's own guidance for App Previews points the same direction: show the app itself, stay inside the app, and don't imply functionality the app doesn't have (Apple App Previews). Google is blunter about it, recommending that at least 80% of a store preview is representative user experience, with core functionality in the first 10 seconds and a hard limit on title cards, logos, and pre-rendered promo material (Google Play preview assets).

Both platforms are telling you the same thing. Your interface is the story. Everything else is packaging.

Four decisions carry the whole video:

Promise, path, proof, placement. If you can't write one sentence for each before you open a screen recorder, the brief isn't ready.

The test for whether a screen earns its place

Founders over-include because every screen took engineering effort. Viewers don't care about effort. Run each candidate screen through this:

1. Does it move the workflow forward, or does it only describe the app?
2. Would the video still make sense if you removed it?
3. Is the text on it readable at phone size without pausing?
4. Does it show something happening, rather than something sitting still?
5. Is it real, staged product data rather than placeholder text?

A screen needs four yes answers to stay. My recommendation is to plan for four to six screens. Below four, the flow tends to feel incomplete. Above six, each screen gets so little time that nothing registers.

Weak version for a budgeting app: home dashboard, settings, category list, a chart, a notification, the profile page. Six screens, zero story. It reads as a product inventory.

Strong version: a receipt photo gets snapped, the amount auto-fills, the user taps one category, the monthly budget bar updates on the dashboard. Four screens, one complete job, visible proof at the end. Same app, same recording session, completely different result.

What to cut, and why it hurts to

Every one of these feels like it belongs. None of them does.

• Long logo reveals and title cards. On a store listing, the preview may autoplay muted in search results, so the first frames are prime real estate. Spending them on branding is expensive.
• Onboarding and account setup. Signup screens show the work required to use the app, not the reason to want it.
• Feature montages. Rapid cuts across unrelated screens look energetic and communicate nothing.
• Unreadable UI. If a data table or settings panel needs a pause to parse, it does not belong at phone size.
• Repeated steps. Three list items being checked off teaches the same thing as one.
• Decorative device spins and 3D phone frames. They eat seconds and shrink your actual interface.
• Any claim you can't show. Apple's asset guidance says to skip prices, discounts, website URLs, unverifiable claims, and other marketplace logos in App Store assets entirely (Apple asset best practices).

One more, and it matters legally as well as creatively: don't fake a feature through editing. Cutting from a tap straight to a finished result implies an automation that may not exist. Apple calls this out directly. If the step takes six seconds in the product, speed it up honestly rather than transitioning past it.

The placement table

Same footage, different jobs. This is the part most teams skip.

Store rules are real constraints, not preferences. Per the Apple preview guidance linked above, you get up to three app previews per supported language, each up to 30 seconds, and previews play muted by default with autoplay that may be disabled, which is why the poster frame matters. Google Play's preview asset rules, also linked above, take one YouTube URL, which must be public or unlisted, embeddable, not age-restricted, with monetization and ads turned off.

For Meta, TikTok, and YouTube ad placements, confirm current specs inside your ad account before export. They change, and a stale spec sheet costs you a re-edit.

Three cuts from one capture

Record once, at high quality, with clean takes of each step and a few extra seconds of handle on both ends of every action. Then derive.

Store cut, roughly 25 seconds:

• 0 to 3s: the app open on the screen where the job starts, no title card
• 3 to 8s: the first meaningful action, clearly visible
• 8 to 18s: the rest of the workflow, one step per beat
• 18 to 25s: the completed state, held long enough to read

Paid social cut, 15 seconds:

• 0 to 2s: the strongest visual moment in the entire recording, usually the payoff or the single most satisfying interaction
• 2 to 10s: a compressed version of the path, two steps maximum
• 10 to 15s: result, then a direct install CTA on screen

The paid cut is not the store cut trimmed from the front. It is reordered. Store viewers already chose to look at you, so you can build. Social viewers haven't, so you lead with the most arresting frame you own. Our breakdown of video hooks that stop the scroll goes deeper on picking that opening frame, and if you're also sizing vertical exports, the Instagram Reels dimensions guide covers safe zones so your UI overlays don't sit under the interface.

Landing page and YouTube cut, 45 to 60 seconds:

• Same workflow, slower pacing
• Room for a second use case or an edge case
• Optional voiceover explaining why the workflow is built that way
• CTA matched to the page, not to the store

These timings are production choices, not industry standards. Adjust them to how long your actual workflow takes.

Capture details that separate usable footage from re-shoots

The difference between a promo video that looks professional and one that looks improvised is almost never the edit. It's the recording session.

Stage the data before you hit record. Real-looking names, sensible amounts, a populated history. An empty state with "Item 1" and "$0.00" makes a funded product look like a prototype.

Turn on do not disturb and clear the notification shade. A banner from a messaging app in the middle of your hero shot means recording the whole sequence again.

Make touches visible. Enable tap indicators or record with a device that shows them, and move deliberately. Fast, confident taps read as random flashes on a small screen. Slightly slower movement with a half-second pause on each destination screen reads as competence.

Disclose requirements. Apple's preview guidance asks that paid content, subscriptions, and login requirements be made clear in previews. A one-line overlay saying subscription required is better than an install followed by an uninstall.

Review the final export on an actual phone, at arm's length, with sound off, outdoors if you can. Text that was comfortable on a 27-inch monitor often disappears. This single check catches more problems than any other step.

Then test the variants against each other rather than arguing about which one is best. A structured approach to ad creative testing replaces review-room opinion with performance data.

Brief readiness check

Before anyone records, you should have written answers to these:

• The one workflow being shown, named as a sentence: "user photographs a receipt and it lands categorized in their budget."
• The four to six screens, in order.
• The completed state that counts as proof.
• Which placements you're cutting for, and the CTA for each.
• Whether a subscription, login, or paid tier gates any step shown.
• Who reviews the export on a physical device.

Missing two or more means you are not ready. Record anyway and you will pay for it in re-shoots.

Producing it yourself or hiring out

In-house works when the workflow is simple, someone on the team has a decent eye, and you only need the store cut. It stops working when you need multiple campaign-ready variants per month, with captions, hooks, and formats per placement. That volume is where internal production quietly stalls.

If that is the point you're at, Viralix is a marketplace where brands and agencies hire vetted AI video creators for campaign-ready video ad packages, with defined scope, revisions, and rights. Useful when you need variants on a schedule rather than one hero asset. If your need is closer to a full product walkthrough than a promo cut, our guide to software demo video production covers that format instead.

Do this next

1. Write the promise, path, proof, and placement sentences for your app. Four sentences, one sitting.
2. List every screen you want to include, run the five-question test, and delete anything scoring below four.
3. Stage your demo data and record the surviving workflow in one clean session with handles on every action.
4. Cut the store version first, then rebuild the 15-second paid version starting from your strongest frame, then extend the landing page version.
5. Check every export on a real phone with sound off before it goes anywhere near an ad account or a store listing.