Sampling distribution

When population variance is unknown, it is mentioned to use the standard error= s/√n, where n is size of sample and s is calculated from s=(xi-x)^2/n-1, where n is sample size and i limit is 1 to n, x is mean of sample means . I don’t understand the logic of calculating standard deviation and variance again if we are already calculating variance of the means of sample mean. Can someone please elaborate if I am wrong in understanding the concept for variance calculation.