The difference in means might be a useful measure of effect size if you are interested in comparing the results of two experiments, each with large n. It tells

The difference in means might be a useful measure of effect size if you are interested in comparing the results of two experiments, each with large n. It tells you what difference to expect if you repeated those experiments with new samples. If, however, you are interested in the effect on a single observation, then the variability in samples/populations needs to be taken into account.