App Preview Video Best Practices: 15 Seconds That Convert (2026)
TL;DR. A good app preview earns its keep by front-loading the hook. Average watch time is only 4 to 6.5 seconds (StoreMaven), and completion is under 20%, so your strongest feature must land in the first 3 to 5 seconds. Apple allows 15 to 30 seconds, up to 3 previews per localization, captured from real in-app footage and autoplayed muted. That means captions over narration, a deliberate poster frame, and a hook before anything else. This is the how-to-make-a-good-one guide. If you are still deciding whether to make one at all, start with the vs-screenshots decision.
Last updated: 2026-07-27.
Disclosure. We make ScreenFast, an AI App Store screenshot generator. It does not make videos, so its role here is narrow and we name it once. Every Apple spec is cited to developer.apple.com, and conflicting conversion stats are flagged.
Why 15 seconds is the number that matters
Apple permits 15 to 30 seconds. The reason to aim short is behavioral, not a rule.
Average watch time across categories is only 4 to 6.5 seconds, and completion runs under 20% (StoreMaven). Apptamin estimates a 4 to 12 second average view. Apple and SplitMetrics both treat a view longer than 3 seconds as a valid view. So most people never reach your ending. Your hero feature has to appear in the first 3 to 5 seconds, and a tight 15-to-22-second cut usually beats a full 30.
To be clear: "15 seconds" is a best-practice target, not an Apple limit. Apple allows up to 30. You are choosing brevity because attention is scarce.
Apple specs and review rules you cannot skip
From Apple's App Previews page and the App Store Connect reference:
- Length and count: 15 to 30 seconds, up to 3 previews per localization, max 500 MB each.
- Format: .mov, .m4v, or .mp4, H.264 around 10 to 12 Mbps, 30 fps, AAC stereo audio.
- Real footage only: capture from inside the app on a real device. You may not film people, fingers, or over-the-shoulder angles. Touch hotspots and graphic elements are allowed to show interaction.
- Content rules: no prices, no seasonal references, no unlicensed music or trademarks.
- Per device family: separate previews for iPhone, iPad, Mac, and Apple TV. Cited device resolutions include iPhone 6.9-inch at 1290x2796 and iPad Pro 13-inch at 2064x2752.
Common rejection reasons: non-app content, showing features the app lacks, wrong dimensions, and playback failures.
The hook: structure your first 3 seconds
Lead with the hero feature or the final outcome. Never open on a logo, splash screen, login, or a slow fade. The first 3 seconds carry disproportionate weight (AppFollow).
Two competing structures from the field, both vendor heuristics, not Apple rules:
- AppScreenshotStudio: Hook/Problem 0 to 5s, Solution 5 to 15s, Benefit 15 to 25s, CTA 25 to 30s.
- SEM Nexus: Hook 0 to 7s, Proof 7 to 15s, with 18 to 22 seconds often beating a full 30.
Whichever you use, cut every 1.5 to 2 seconds. A muted, short-form feed punishes static shots.
How to Create the Best App Preview Video walks through the hook and edit on screen.
Design for muted autoplay: captions over narration
App previews autoplay muted on the product page and in search results (Apple). Roughly 80% of users watch with sound off. So the message has to work in silence.
Use large, high-contrast on-screen captions, visible touch indicators, and motion instead of voiceover. As AppFollow puts it: if your entire message depends on a voiceover, most users will miss it. Voiceover is optional and often not worth the cost. Budget caption legibility first. Keep the typography and color consistent with your screenshot set so the listing reads as one cohesive story, which is what Apple recommends. The caption craft carries over from screenshot copywriting.
The poster frame: your second-most-important screenshot
The poster frame is the static image shown before playback, and the one shown to users with autoplay off. Apple defaults to the frame at about the 5-second mark unless you override it, and that default is often weak.
Treat the poster frame like a screenshot. Pick a moment with clear UI, legible text, and visual consistency with screenshot 1. Even if you do not override it, make sure the 5-second mark contains a strong frame. Changing the poster frame on an approved preview requires resubmission, so decide deliberately. This is the one spot where a screenshot tool helps: the poster frame is a still image, and ScreenFast (our product, which makes screenshots not videos) can design that frame even though it cannot produce the video itself. See the AI App Store screenshot generator.
How to capture footage: real device or simulator
Apple's recommended tools are QuickTime Player, iMovie, or Final Cut Pro X. In QuickTime: File, then New Movie Recording, then select the connected iPhone. Export via File, Share, App Preview. Capture from the exact device family you target; do not resize footage from another model.
No physical device? Use the iOS Simulator. File, then Record Screen (Cmd+R), or xcrun simctl for custom resolution, codec, and automation. Capture at native UI resolution. App Previews require 30 fps, so confirm your capture and export hit that frame rate. The capture mechanics overlap with the simulator screenshot workflow.
Does a preview video actually lift conversion?
Honestly, less than the marketing claims suggest, and it depends on category.
The most defensible figure is SplitMetrics, which measured roughly a 16% average conversion increase from optimizing video previews. AppFollow's per-category lifts are more modest: Gaming +8 to 18%, Health and Fitness +7 to 14%, Fintech +6 to 12%, Social +5 to 11%, E-commerce +4 to 10%, Utilities +3 to 9%. The widely repeated +20 to 40% (and a 47% AppTweak figure) lacks a fresh primary study, so treat it as a ceiling, not an average.
And the downside is real: a poorly executed video can drop conversion 5 to 15% (SEM Nexus). A bad video is worse than no video. The category data lives in our conversion benchmarks.
How to A/B test your preview
Use Apple Product Page Optimization (PPO) and Custom Product Pages (up to 70 per app since Oct 2025). Test video-versus-screenshots-only, opening scenes, caption and feature order, and value props.
Practical thresholds from the field: target roughly 1,000 installs per variant minimum (AppFollow), run for at least 14 days, and ship a winner only at a 3% lift or more (within 2% means keep iterating, per SEM Nexus). Measure download-to-paid and revenue per download, not installs alone (AppScreenshotStudio); Apple's WWDC25 analytics added download-to-paid tracking. The common variables to test: length, voiceover on or off, on-screen text, music, and poster frame.
| Variable | Apple rule / spec | Best practice |
|---|---|---|
| Length | 15 to 30s, up to 3 per localization | Front-load the hook; 15 to 22s |
| Audio | Autoplays muted; AAC stereo if used | Design for silence, captions over voiceover |
| Footage | Real in-app capture on target device | Real workflows, not stock or marketing |
| Poster frame | Defaults to ~5s frame; override needs resubmission | Pick a screenshot-quality frame deliberately |
| Format / fps | .mov/.m4v/.mp4, H.264, 30 fps, max 500 MB | Capture and export at 30 fps |
| Testing | PPO and Custom Product Pages, up to 70 per app | ~1,000 installs/variant, 14-day run |
How we tested
All hard specs are cited to Apple's App Store Connect reference and the App Previews page, not paraphrased from vendor blogs. Every conversion statistic is attributed to its named source with its bias flagged (StoreMaven, SplitMetrics, AppFollow, AppTweak, Apptamin, SEM Nexus). Where claims conflict, we preferred the conservative, dataset-backed figure (SplitMetrics ~16%) over the unsourced 20 to 40% and labeled the latter a ceiling. This is the how-to-produce guide; the if-you-should-make-one decision is covered in our preview video vs screenshots article. ScreenFast is our own product, mentioned only at the poster-frame step it can actually serve.
FAQ
What length should an app preview video be: 15 or 30 seconds?
Apple allows up to 30, but aim for 15 to 22. Average watch time is only 4 to 6.5 seconds, so a tight cut with the hook up front usually outperforms a full 30-second video.
Do I need a voiceover if the video plays muted?
No. Previews autoplay muted and about 80% of users keep sound off. Spend your budget on large, legible captions and clear motion, not narration.
Can I capture an app preview without a physical iPhone?
Yes. Use the iOS Simulator's Record Screen (Cmd+R) or xcrun simctl, capturing at native UI resolution and 30 fps. QuickTime with a connected device is Apple's other recommended path.
How much does a preview video actually lift conversion?
The defensible average is around 16% (SplitMetrics), with per-category lifts of roughly 3 to 18% (AppFollow). The widely-quoted 20 to 40% is a ceiling claim. A poorly made video can also lower conversion by 5 to 15%.
What is the poster frame and how do I choose it?
It is the static image shown before playback or when autoplay is off. Apple defaults to the 5-second mark. Override it with a screenshot-quality frame that has clear UI and matches your first screenshot.
How do I A/B test my app preview video?
Run it through Apple Product Page Optimization with about 1,000 installs per variant over at least 14 days, and ship a winner at a 3%+ lift. Test the opening scene, length, captions, and poster frame.
Should I use the same style as my screenshots in the video?
Yes. Apple advises a cohesive story. Match the typography, color, and tone so the video and screenshots read as one listing, not two disconnected assets.