Once the test is complete, analyze the data to determine which version performed better. Look at the metrics you decided to measure and compare the results against your hypothesis. Use statistical significance calculators to ensure that your results are not due to random chance.