p-value is the probability that the trial results you observe could happen strictly by chance if there was no difference between the treatments. Very small p-value means very small chance that results are due to chance, so effect must be due to the treatment.