The Biggest Mistake People Make When Split Testing

Split testing is one of the most important things a website can do. It allows you to test different designs and content to see how they do against each other. In essence, a split test gives you the chance to test out new designs to see which one will get you the most conversions. Doing a split test is fairly simple, but there is one mistake that most website owners make. They just don’t wait long enough for the test to really get a good amount of information.

The Damage of Impatience

The majority of website owners known split testing is necessary, but most don’t wait long enough. They think they can make an informed decision after gaining some simple information. That’s just not the case.

Being impatient can really damage your website, and kill your conversion. Think about this for a moment. You test two different designs. Within one week, you notice that your original design is raking in better sales, so you abandon the new design and stick with the old one.

If you tested longer, then you might have found out that the other design could actually pull in 20 percent better conversions. That can be a lot of lost money on the table. You need to keep testing until you have enough information.

Longer Tests

You know why longer tests are needed, but why would a short test be inaccurate? A short test only gauges the results of a few sales. When a business is about to release a new product, it doesn’t just ask a few people what they think about the product. The business goes after hundreds or even thousands of people. Larger tests give you a better picture of what your entire demographic thinks, not just a small portion of it.

By performing a short test, you are just relying on information from a small portion of your market. If you want to really improve your website, then you cannot use short tests. You need to split test until you get a good amount of statistically relevant data.

Optimum Test Length

Just how long should you let the test run? That is a hard one to answer. It matters on how many current customers you have, how long your website has been doing business and many other factors.

For example, the Amazon website would probably wait several months and after thousands or even millions of sales before deciding between two different designs. Your website probably isn’t that large, so you probably don’t need the same amount of information.

You usually want to wait over a month for the split test to really capture data, though larger websites should wait at least several months. You also want to capture a good amount of sales. Relying on information from just 10 percent of your monthly sales would be a bad idea.

Let at least one hundred sales occur, though most websites should actually wait for more. Just let the sales keep going until you feel you have enough information to make a good decision. Leaving money on the table is never a good idea, so let customers come in and keep buying until you see one design really win over the other.


Split testing is a great way to increase your sales. It gives you the chance to test different designs, and it gives you the chance to improve your conversion. Just make sure that you let the test run its full course. If you don’t, then there is really no point in performing the test in the first place. A short test is an inaccurate one, so let it run until you have enough information.