A system and method track short term and long term intervals to assess whether a voice message source is a likely source of voice spam. Upon detection of a spamming threshold, calls from the source are blocked until detection that a sufficient time interval has elapsed without generation of messages.