The playable end card web shop link, done properly

Since Judge Yvonne Gonzalez Rogers held Apple in contempt on 30 April, ruling that the company cannot charge commission on external link purchases or restrict how those links appear to US users, playable teams have been getting the same request from growth leads: put a web shop link on the end card. The ruling made the commercial case simple. It did nothing for the production work, and most of the playable end cards going out right now treat the web shop link as a decoration bolted onto a template built for something else.

Why the end card is harder than it looks

A standard playable end card does one job: turn a completed session into an app store click, with a single CTA wired to a store listing and a fallback that reopens the store page when a deep link fails. A web shop link doubles the number of destinations the end card has to route to. It adds a second CTA that has to read as distinct without competing with the primary install button, and it points at a target URL that is not a store listing and therefore does not arrive pre-trusted by every network's creative review process. None of that is a design problem so much as a plumbing one. Networks are handling it inconsistently because none of them had a mature spec for it before the end of April.

What has to be true before the link goes live

A web shop link on a playable end card only works if four separate things hold at once, and treating them as one problem is where early implementations have gone wrong.

• The destination URL is device- and platform-aware, sending an Android user somewhere different from an iOS user, since Google's checkout rules are not the ones under contest in the Epic v Apple case.
• The link opens outside the network's in-app browser wherever the SDK genuinely allows an external open, because a web shop rendered inside a webview can inherit the app's payment context in ways that reintroduce the commission dispute the link was meant to avoid.
• The click carries its own attribution parameters, separate from the install click, because a measurement partner that only tracks the install path on a two-destination end card will under-report the web shop path entirely.
• The fallback state, what a user sees if the web shop link fails to open, does not silently drop them back to the install CTA, since that would erase the entire point of offering the option.

A QA checklist before anything ships

Playable QA has always been about catching the failure that shows up on exactly one device or one network's rendering engine, and a second CTA doubles the number of ways that can happen. Before a web shop end card variant goes out:

• Test the web shop CTA on the oldest supported OS version in the spec, not just the newest, since external link behaviour on iOS has varied by version historically.
• Confirm the web shop link opens correctly inside every network's own preview or test harness, not just in a browser, because several networks render playables inside a constrained webview that does not behave like Safari or Chrome.
• Verify that closing the web shop tab returns the user to a live end card rather than a blank state, so a user who backs out still sees the install CTA.
• Check that the web shop click event actually fires and reaches the MMP before assuming any conversion data from the new destination is trustworthy.
• Confirm any required compliance copy for the market is present on the web shop CTA itself, not just buried on the landing page it opens.

Why networks are handling this inconsistently

Part of the mess is structural rather than lazy. Every network that renders playable ads runs its own creative review pipeline, its own webview implementation, its own rules about what a click can point to, and nobody drew up those pipelines with a second, non-store CTA in mind. A network built around a single install-now click event has to add a second event type, decide how it counts toward billing and optimisation, then update its creative review checklist, all inside a few weeks of a court ruling nobody had scheduled for. Some have handled this by allowing the web shop link only as a static end-card element with no interaction tracking at all, which satisfies the letter of "the link exists" while providing none of the attribution a growth team actually needs. Treat every network's current web shop support as provisional until its own documentation catches up. Confirm directly rather than assuming parity across two networks that both claim to support the feature.

What is different market by market

The contempt ruling applies to US iOS users specifically, which means a global playable spec now needs a geo condition it did not need in March: US iOS traffic gets the web shop CTA, everyone else gets the standard single-CTA end card, at least until other markets change their own rules. Fortnite's resubmission to the US App Store this week is a reminder of how contested this territory still is even for a company with Epic's legal resources. A smaller studio's playable spec should assume the rules around what a web shop link can say, and how prominently it can appear, will keep shifting through the rest of the year rather than settling now.

Building the spec so it survives the next change

The studios handling this well are not the ones who shipped a web shop CTA fastest. They are the ones who built the geo condition, the platform-aware routing, the separate attribution tagging, all of it as a configurable layer in their end card template rather than as a one-off build for a single campaign. As this column argued in "External purchase links on iOS: what to build first," the technical foundation for a compliant web shop flow and the foundation for a fast-shipping playable spec are the same foundation: build it once, reuse it. A playable end card that treats the web shop link as a permanent, configurable option rather than a temporary bolt-on will need far less rework when a court ruling, a platform policy change or a new market's link-out rules moves the goalposts again.

Related archive reading

These articles provide related context and remain subject to their stated review status.

• What web shop conversion rate should you actually expect
• The Web Shop Attribution Blind Spot Nobody Budgets For