The standard deviation (SD) in fluorescence correlation spectroscopy (FCS) has been mostly neglected in applications. However, the knowledge of the correct SD is necessary for an accurate data evaluation, especially when fitting theoretical models to experimental data. In this work, an algorithm is presented that considers the essential features of FCS. It allows prediction of the performance of FCS measurements in various cases, which is important for finding optimal experimental conditions. The program calculates the SD of the experimental autocorrelation function online. This procedure leads to improved parameter estimation, compared to currently used theoretical approximations for the SD. Three methods for the calculation of the SD are presented and compared to earlier analytical solutions (D. E. Koppel. 1974. Phys. Rev. A. 10:1938-1945.), calculation directly from fluorescence intensity values, by averaging several FCS measurements, or by dividing one measurement into a set of shorter data packages. Although the averaging over several measurements yields accurate estimates for the SD, the other two methods are considerably less time consuming, can be run online, and yield comparable results.