Online reviews are very important for any business that wants to control its online reputation. This allows businesses to have active and positive participation from consumers. As an information and communication company in Indonesia PT Telekomunikasi Indonesia Tbk commonly called Telkom require a customer's perspective or review to maintain the relevance of their digital products on the market. One method often used to analyze online reviews is sentiment analysis. Sentiment Analysis is used to gain an understanding of the opinions, attitudes, and emotions expressed in the mention of online by determining the emotional tone behind a series of words.This research tries to compare classifications in sentiment analysis of Telkom's product from consumer reviews written in the form of tweets on Twitter. Each tweet about Telkom digital products such as Indihome, UseeTV, and Wifi.id will be collected as data. The use of classification types will be compared to help with the accuracy of sentiment analysis based on three types of methods TextBlob, Naïve Bayes & K-NN (K-Nearest Neighbor).The best result of this research is the K-NN algorithm with an accuracy score of 75% followed by Naïve Bayes 69.44% and the last is TextBolb with 54.67%.