Most app teams treat App Clips and Android Instant Apps as a 2019 experiment that never went mainstream. That assumption is keeping them from the one pre-install surface where intent is genuinely high and attribution is genuinely possible.
In the right hands, an App Clip or an Instant App turns a QR code at the till, an NFC tap at the door, or a Google Search result into a working transaction in seconds, no install required. In the wrong hands, it is a six-figure engineering project that no marketer ever uses. This playbook covers what they actually are, the categories where they earn their keep, how attribution survives the upgrade to a full install, and the six-week pilot plan that lets you find out without committing the roadmap.
What App Clips and Android Instant Apps Actually Are
An App Clip is a lightweight, on-demand slice of an iOS app, launched from a URL, NFC tap, QR code, App Clip Card, or Maps, that lets a user complete one focused task without installing the full app. Android's equivalent is an Instant App, launched through verified App Links from Google Search or Play Store. Apple's overview page at developer.apple.com/app-clips and Google's at developer.android.com/topic/google-play-instant are the canonical references.
A few specifics most marketer-facing posts leave out:
- App Clips ship in a small size envelope (around 10 to 15 MB depending on launch surface, per Apple's public spec), which forces a single-purpose design.
- Instant Apps run from a verified App Links setup, so the assetlinks.json file you ship for deep linking is the same file the Instant App entry point depends on.
- Neither surface replaces a full install. Both are designed to make the first useful interaction faster, then route the user toward the full app if they want more.
For a refresher on how the wider deep linking family fits together, our complete guide to deep links for app teams in 2026 covers URI schemes, Universal Links, App Links, deferred, and dynamic links in one place.
Where They Genuinely Fit (And Where Marketers Should Not Bother)
App Clips and Instant Apps were built for transactional moments where the user wants to act, not browse. The categories that get a real payoff have a few things in common: a clear single-task flow, a physical or short-form entry point, and a payment or booking action at the end.
Where they earn their keep:
- Food, transport, ticketing, parking, retail. The user arrives with intent (QR at the table, NFC at the gate, Maps tap at the cafe), and the App Clip handles a single transaction.
- One-time payments and bookings. The user does not need an account history to act.
- Offline-to-app handoffs. QR codes and short-form entry points anchor the flow, which links neatly to the patterns in our deep linking guide for QR codes.
Where the payoff is marginal:
- Subscription apps and content apps. The user needs a logged-in profile and a longer relationship; an App Clip cannot represent that meaningfully.
- Social and community apps. The first useful interaction is a feed, not a task.
- Games. The size envelope and the single-task design fight against everything games are trying to do.
If your app does not have a single-step transactional surface that already lives in the wild on a physical object or a search result, neither App Clip nor Instant App is the right next investment.
How Deep Linking Works Inside an App Clip and Inside an Instant App
App Clips and Instant Apps are deep linking surfaces, not separate apps.
- An App Clip is triggered by an invocation URL configured against your App Clip experience. The same URL pattern that drives your Universal Links is the one that launches the Clip.
- An Android Instant App is triggered through verified App Links. The assetlinks.json hosted at your domain is what lets Google Search and Play Store route a user into the Instant App without an install.
- What carries over to the full install is the deep link context (the URL, its parameters, and any deferred payload your SDK preserves). What does not carry over is the App Clip's local cache, its sandboxed identifiers, and any session state you have not explicitly persisted.
The handoff to the full install is where most attribution chains break, and it is the part of the design most teams underestimate.
The Attribution Question Most Teams Get Wrong
If the App Clip session is the only interaction, attribution looks like any other deferred deep link click: the source, the campaign, and the conversion event sit inside the Clip's reporting.
The harder case is the upgrade. A user who completes a task inside an App Clip and then installs the full app is, on most stock setups, two separate user records: one in the Clip's reporting and one in the full app's reporting. Without a deferred deep link that survives the upgrade, the campaign that drove the original Clip session never gets credit for the install.
Consider a food-delivery app shipping a QR-to-order App Clip at around 200 outlets. The Clip can take payment without an install. But unless the deferred deep link carries the outlet ID and the offer code through to the full install, the campaign that placed the QR sticker never sees the downstream lifetime value. Our explainer on deferred deep links walks through how that handoff is built.
An MMP that does not handle App Clip postbacks treats the Clip session and the full install as disconnected events. If you are evaluating tools, check this against the criteria in our 10 non-negotiable features every deep linking tool must have in 2026 checklist.
The Engineering and QA Cost (Be Honest)
The reason most teams skip App Clips is not a marketing decision; it is a backlog decision. A realistic cost picture, before you commit:
- App Clip Card design, App Clip experience configuration, the App Clip's separate code path. It is a separate binary with its own size budget and its own entitlements.
- App Links verification for the Android side. Assetlinks.json hosting, the domain verification flow, and Play Console configuration for the Instant App entry point.
- QA matrix worth budgeting for. Clip from camera, Clip from Safari, Clip from Maps, Instant from Google Search, Instant from Play Store. The matrix is small but each cell can fail differently. Our pre-launch deep link QA checklist covers the testing discipline that prevents most launch-day surprises.
Realistic ballpark for a single transactional flow shipped on both iOS and Android: around six weeks of focused engineering and QA, with a small product and design slice on top.
A Six-Week Pilot Plan for One Transactional Flow
The right way to find out whether App Clips earn their keep for your app is a single-flow pilot, not a platform-wide rollout.
- Weeks 1 to 2. Pick the highest-intent entry point. For most transactional apps that is a QR at point of sale; for some it is an NFC tag at a venue or a Google Search result for a branded query.
- Weeks 3 to 4. Build the minimum viable App Clip and Instant App for that one flow. Resist the temptation to handle three flows; the engineering cost compounds.
- Weeks 5 to 6. Instrument attribution end to end. Push the Clip to a small live audience. Measure two ratios: the Clip-to-install rate (how often a Clip user upgrades to the full app), and the Clip-to-revenue rate (how often a Clip session ends in payment).
After six weeks you have two numbers that decide the next move. If both ratios clear your internal thresholds, you have a wedge worth scaling. If neither does, you have spent six weeks instead of six months learning that this surface is not for you.
FAQ
What is the difference between an App Clip and a Progressive Web App?
An App Clip is a sandboxed slice of a native iOS app, signed by Apple and launched without an install. A Progressive Web App runs in the browser. The Clip has access to native APIs and pays the size cost; the PWA has neither.
Can an App Clip pass campaign attribution to the full install if the user upgrades?
Yes, but only if the deferred deep link is wired through the upgrade. The URL that launched the Clip needs to be preserved and surfaced to the SDK on first open of the full app. Without that bridge, the install looks organic.
Do Android Instant Apps still work in 2026?
Yes. The Google Play Instant programme is still available and depends on the same App Links verification you use for deep linking. Coverage is narrower than App Clips in practice because Google has invested less in launch surfaces beyond Search and Play Store.
What is the size limit for an App Clip, and has it changed?
Apple's App Clip envelope sits around 10 to 15 MB depending on the launch surface, per the App Clips developer documentation. Treat the smaller figure as the planning constraint to avoid edge cases on the lighter launch surfaces.
Which app categories see the highest payoff from shipping an App Clip?
Apps where the first useful interaction is a single transaction tied to a physical or short-form entry point: food ordering, transport, ticketing, parking, retail. Apps with longer relationship-building flows see far less return.
Where to Go From Here
App Clips and Android Instant Apps are not a marketing surface for every app. They are an unfair advantage for the small set of apps where the first useful interaction is a single transaction at a physical or short-form entry point. The six-week pilot is the cheapest way to find out which set yours sits in.
If you want to instrument the Clip-to-install handoff cleanly on your own apps, request a Linkrunner demo and we will walk through the deferred deep link setup that keeps campaign context intact across the upgrade. Pair it with the QA checklist above and you have a working pilot in a single sprint cycle.