The disclosed embodiments relate to methods, devices and computer program products that enable watermark-based metadata recovery for content with multiple alternative components. The content can, for example, be included in broadcast television services which include multiple alternative audio and video components, only a subset of which may be presented by a given receiver at a given time. The disclosed techniques, among other features, enable broadcasters to deliver different services to different customers by using different watermarks into the feeds they provide to multichannel video program distributors (MVPDs). For instance, coincident audio and video components are embedded with different watermark payloads.