A/B testing is good, but not perfect for everyone...

A/B or multivariate testing is used by marketers for discovering the best variation of the landing pages. But if you want to optimize fast, A/B testing might not be the best choice.

• Much traffic and time to achieve statistical significance.A/B testing takes thousands of your landing page visitors and months to achieve 95% statistical significance if you don't have huge traffic.
• You can only test one variable.With A/B testing you can usually test only one thing at a time: headline, your call to action, minor layout change and etc. If you change multiple things - it's multivariate testing and it takes even more time than A/B testing.
• You need to choose what to test.Your A/B tests could fail because you chose the wrong things to test. It could be that you choose things that don't actually change your conversion rate in any measurable way.
• Sending visitors to an underperforming landing page.You send your landing page visitors to different variations of your landing page regardless of how they’re converting. It happens that your potential customers get an underperforming page for weeks while you're waiting for the test results.
• Bad segmentation and personalization.A/B testing has no context about your visitors. Your visitors are different. It depends on how they arrived on your landing page. What are they looking for? Which country do they come from?
• Loading speed and flickering.A lot of A/B testing tools have some side-effects on your landing page loading speed. It happens that the original variation content gets briefly displayed before applying a different variation.