The fitting of regression model has problems related to non-linearity, multicollinearity, serial correlation and heteroscedasticity which involves very long and complex procedure of calculations and analysis. This study focuses on an improvement in the model fit based on R² value. An attempt is made to investigate the outliers in any data set and to increase the R² square value after the removal of outliers. In this study, a hypothetical data set is considered.  The data set indicates consumption as a dependent variable and Income, Food size are considered as independent variables. The regression model for Actual data indicates the R² value is 0.455. After the removal of outliers using the cook’s distance, the revised R² value is 0.578. This indicates that the outlier in the data set plays a vital role in the model fit. Therefore it is necessary to remove the outlier if any in the data, before proceeding to further analysis.