Sometimes a Unicode character such as a smart quote ( 〞), a long dash (—), or Unicode whitespace accidentally slips into your carefully crafted 125 character message and now your message is segmented and priced at two messages instead of one.
When Unicode characters are used in an SMS message, they must be encoded as UCS2. UCS2 characters take 16 bits to encode though, so when a message includes a Unicode character, it will be segmented at 70 characters instead of 160 characters per message segment.
The Messaging Service Smart Encoding feature helps detect those easy to miss Unicode characters and replaces them with a similar GSM encoded character. This helps ensure that your message gets segmented at 160 characters and saves you from sending two messages.
Messages that include emoji or language-based characters such as kanji are not transliterated with Smart Encoding. Please reference this list for the Unicode characters Smart Encoding will replace.