A Penalty analysis was used to identify potential directions for the beverage’ improvement, based on consumer tests. Six samples of coffee, black tea and red tea were prepared, half sweetened using sucrose and the other half with stevioside. The acceptance analysis used a nine-centimeter unstructured scale, anchored with the terms (1=extremely dislike; 9=extremely like), in relation to appearance, aroma, flavor, texture and overall liking. The samples were presented to 120 judges, untrained, in monadic form and using a complete balanced blocks design. The same group analyzed the samples using a 5-point Just About scale (JAR) regarding sweetness and bitterness. The Fizz Sensory Software model 2.40H was used to collected data. Effective results were calculated by analysis of variance (ANOVA) and Tukey’s means test (at 5% significance level). The percentage of respondents scoring in each category was calculated for each sample. The respondent percentages (x-axis) were plotted against the penalties (y-axis), and an attribute was considered significant when the respondent percentage was higher than 20% and the penalty score (drop in overall liking) was higher than 1. Penalty analysis based on JAR scales in relation to red tea and coffee showed that bitterness was the most penalizing sensory characteristics causing significant drop in these beverages acceptability. Among the evaluated attributes, sweetness was the more significantly responsible for the penalty analysis of black tea acceptance. From affective results, there was no significant difference between black teas, sweetened with sucrose and stevioside. For red tea and coffee, consumers perceived significant difference in relation to flavor and only for coffee, the overall liking was significant different. The black tea, a traditional bitter beverage, showed that stevioside did not affect consumer perception. Thus, there was an influence between products originally considered bitter when then were sweetened using stevioside.