Here are the steps of the kMeans clustering algorithm, put them in order.

Step $1$:

Once all points have been allocated, compute the new cluster centroid

Step $2$:

if the centroid positions have changed significantly, go back to step $2$

Step $3$:

Loop over all of the data points

Step $4$:

Select K random points as initial centroids

Step $5$:

Allocate the point to the closest cluster centroid

Step $6$:

Calculate the distance from the point to each cluster centroid