A digital answering machine, electronic voice mail system, or any other voice messaging system integrating both non-speech data such as caller ID information or graphical or pictorial images together with associated speech data in common memory such that only one message table is necessary. The integrated common memory is divided into sectors, and each sector is divided into pages which each contain both speech and non-speech data. A single message table is associated with each corresponding set of speech and non-speech data, and includes an offset value indicating the beginning address of the speech data stored just above its associated non-speech data in the first linked page. The offset value is variable to provide an elastic non-speech data buffer to accommodate variable length non-speech data such as would be required by the various caller ID standards throughout the world.