Our Blog

5 Do’s and Don’ts of Split Testing


If you’ve decided to run a split test to optimize the performance of your page, you want to make sure you do so correctly. The best practices will help you achieve your desired results, while common mistakes can keep you from increasing your coversions

1. Don’t rush your tests

Oftentimes, people wonder how long an appropriate time is to run a split test. Approximately two to four weeks should give you enough time to give you consistent data. If you’re strapped for time, you might be satisfied with one week, but at least two is optimal. Keep in mind that your users are going to act different during different times of the day, so having enough time to gather data is going to be really important to drawing conclusions about what is and isn’t working.

2. Do focus on conversions

Remember that conversions rather than amount of traffic are key to split testing timelines. At least 100 conversions per variant should be the minimum for determining the outcome of your tests.

3. Don’t test multiple variables at once

Of course, it’ll take more time to complete your split testing when you’re just testing one variable at a time, but this is the most effective method. Testing one variable at a time takes the guesswork out of what is leading to conversions and what is not.

4. Do test continuously

It’s important to keep in mind that there’s always room for improvement. Split testing should be part of your routine in helping you better understand your audience. Keep doing it!

5. Don’t ignore the results

When you’re done with your analysis, you want to make sure that you act on your discoveries. Implement changes if your results were statistically significant, and use what you’ve learned as you move forward.