Voice interaction between users and IT devices has become possible through voice user interface. Since VUI is based on vocabulary language, the most basic form of human communication, it lowers the entry barrier for those who experience difficulty using graphic user interface (GUI) or situations where using IoT devices with tiny screen or no screen at all. User experience can be a critical marker for the development of technology, however, few academic researchers have attempted to investigate the users' speech behavior pattern when they use VUI. Therefore, this study aims to provide insight about guidelines and design feedback for the novice users through observation in their adaption process of VUI. The author investigated speech patterns of error correction through observing 'iPhone Siri' users who have less experience on VUI. Through the 7 days Siri usage history, the speech behavior patterns of error correction were analyzed in tone (amplitude, pitch, and duration), and in sentences (sentence elements or structure) based on the Communication Accommodation Theory. The results of this study can be summarized as follow; 1) After the first error, users change their speech tone and sentences at a similar rate, 2) but, when the second error occurred, they change sentences more, 3) the speech behavior patterns after error were different depending on the types of errors and 4) the users' speech behavior patterns were changed depending on time period of use. This study has primarily focused on variation in use behavior by observing real user experience. Through demonstrating how users respond to the situations of errors, this research has practical implications for the industry as it can positively contribute to the development of VUI voice feedback design and improve user experience on VUI.