Using random selection can help marketers ensure that their test results are statistically significant and not skewed by any biases. It ensures that every subscriber has an equal chance of being chosen, which makes the results more reliable. This is particularly important in large-scale campaigns where even small biases can lead to incorrect conclusions.