It may have to do that often actions on set top boxes are related to accessing content, for which the software has to check permissions; these are usually embedded in the stream (think mumultiplexed), and the decoder needs to receive the frames before being able to extract the content and display it. Cable network has very high latency (compared to fiber or adsl).

If some of the credential checks are stored in a separate device (dongle, smartcard ), then it could take even longer.