This paper aims to establish an efficient model for predicting company sales by leveraging logistic regression's strengths. A real dataset of Citrix used to figure out a significant variable affecting sales acceleration and to find an appropriate metric to measure unstructured information. To build an efficient model, we used two statistical methods; logistic regression and linear discriminant analysis. The classification accuracy of the models compared using Fisher Analysis, ROC curves, and confusion matrix. In regression analysis, it is evident that response and predictors some times may suffer from correlation issues. By definition, Multicollinearity is that two or more predictors are correlated; if this happens, the coefficients' standard error will increase. Increased standard errors mean that the coefficients for some or all independent variables may be significantly different from 0. In other words, Multicollinearity makes some variables statistically insignificant by overinflating the standard errors when they should be significant. In this paper, we concentrate on logistic regression analysis, linear discriminant analysis, Multicollinearity, fisher analysis, and consequences and effects on the reliability of the regression model