Deciding whether a new capability belongs in the operating system or in an app

You are a product manager at a company that ships an operating system and also builds apps that run on it. New capability proposals arrive regularly, and for each one the company has to choose where it lives: built into the operating system, or delivered at the application level. Explain how you would make that call. Give the criteria you would use, how you would weigh them against each other, and show the reasoning on at least one concrete capability. ```hint Follow the dependencies For any proposed capability, ask who else needs it, which parts of the device or of other apps' data it has to touch, and how often it will need to change after launch. Each answer pulls the decision in a direction. ``` ### Clarifying Questions - Which operating system is in scope: a phone OS, a laptop or desktop OS, a watch, or a TV platform? Release cadence and the cost of an OS change differ a lot between them. - Does "application level" mean one of our own first-party apps, or leaving the capability to third-party developers building on the platform? - Is the goal to rule on one specific proposal, or to give the platform organization a reusable rule for future proposals? - How quickly do devices in the installed base receive OS updates: within weeks, or do many stay on versions that are several years old? ### What a Strong Answer Covers - Treating placement as a spectrum, including the middle options: an OS-level service or API that apps build on, or a preinstalled system component that updates outside full OS releases. - Criteria rooted in the capability itself: breadth of need across users and apps, privileged access to hardware or cross-app data, security and privacy stakes, and the value of consistent behavior everywhere. - Cost and speed of change: OS release cycles, update adoption, backward-compatibility commitments, and how hard an OS feature is to remove once apps depend on it. - Ecosystem and regulatory effects: whether building it into the OS helps or crowds out third-party developers, and the scrutiny that bundling can attract. - A worked example, plus the evidence that would move a capability from an app into the OS (or back) over time. ### Follow-up Questions - A capability launched inside one of your apps has become very popular. What signals tell you it should move into the OS, and how would you migrate existing users and developers? - Several third-party apps already sell a version of the capability you want to build into the OS. How does that change your decision? - How would your answer change for a phone OS whose updates reach users through device makers and carriers, compared with a desktop OS the company updates directly?