Split Testing 101 – Methods that Work

Split testing is one of those things a lot of marketers neglect to use, because they think it’s overly complicated or isn’t really necessary.

And it’s understandable. It does seem like it would be a bit technical, but it’s actually a lot easier than you probably think!

So what is split testing, anyway?

Split testing, in case you haven’t heard of it, is the process of testing multiple versions of a particular piece of copy in order to find out which one converts best.

For example, you could test multiple versions of a page’s headline, or multiple color variations of a sales page.

You can do this manually by simply running one version for a while and tracking results. Then making a change and allowing it to run for a while and comparing results.

Or by using a special piece of software that shows the different versions to visitors on a rotating basis.

Why Use Split Testing?

Perhaps the biggest argument people present when they try to explain why they aren’t using split testing is that they just don’t think it’s necessary.

Why bother with something that might only increase your conversion rate by a percent or two?

Well, in the beginning, it might not seem like that big of a deal. If you’re only making $100 per month, it might not seem like such a huge thing to your bottom line.

BUT…

Don’t think I’m saying you shouldn’t use split testing just because you aren’t making much, yet!

Just because it won’t add thousands of dollars magically to your income doesn’t mean you shouldn’t use it!

Let me explain why...

First, it could actually double your income with very little effort. Yes, that’s right... I said it could double your income!

The reason I said it might not seem like “that big of a deal” is because doubling your income from $100 to $200 might not seem that big.

Types of Things to Split Test

There are tons of different things you can test with split testing. Most people think of things like headlines and designs, but there are so many elements that can be tested.

Let’s take a look at some examples of the types of elements you could split test. This will vary depending on the type of page you want to test, of course.

Things you should consider split testing:

-Headline (text, color, size, font, etc.)

-Pricing (including base price, upsell prices, payment plans, etc.)

-Article titles (clickbait)

-Call-to-action (type, text, color, placement, etc.)

-Offers (eBooks, videos, webinars, etc.)

-Copy (different elements of your actual copy)

-Form fields (name, email, other information)

-Design of a lead-generating pop-up

-The page as a whole

Things you should not bother to test:

-Minor changes to text

-Minor changes to appearance, such as a change of bullet graphic

-Things visitors can’t always see, such as whether a site is mobile friendly. (Your site should always be mobile friendly!)

It should be fairly instinctual to figure out which elements are most important for you to test on your page.

Whatever you’ve agonized over the most, such as price, headline, or design, is probably the first thing you should test.

Multivariate Testing

A/B testing refers strictly to testing two versions of a single element to see which one performs best. Sometimes you may test three versions of the same element.

Such as the headline, but this is usually only effective when you have a lot of traffic. If you don’t have much traffic.

It’s important to limit the number of variations in order to get an accurate result as quickly as possible.

Multivariate testing is different, because you’re testing multiple elements concurrently. For example, you could test two versions of the headline. Two versions of the price. And three versions of the call-to-action.

Multivariate testing is best done when you have a significant amount of traffic. Because with so many versions to test. It’s difficult to get a statistically significant number without quite a lot of traffic. You’d need to get thousands of hits per day for multivariate testing to truly be effective for you.

In this post, we’re specifically going to focus on A/B testing. Because most people don’t have the level of traffic needed to conduct accurate multivariate tests.

Split Testing Best Practices

Split testing can be a little technical, but as long as you follow the standard set of best practices, you should have little trouble.

Let’s take a look at the most important guidelines to follow when split-testing:

One Test at a Time – If you’re split testing manually, the most important thing to keep in mind is that you should only perform one test at a time. Because you want to be certain which element it was that caused the increase or decrease in conversions. You can perform concurrent tests if you’re using automated software. Because the program will detect which element was served and keep the results separate. On your own, it’s not possible to run multiple tests simultaneously while getting accurate results.

One Variable at a Time – Unless you’re using automated software that lets you test multiple variables in a very controlled way. You should only test a single variable at a time. For example, test two different headlines in a single test. But wait until that test is finished to test two different calls-to-action. Again, automated software will let you test multiple variables at a time. Keep in mind that the more variables you test concurrently, the longer it will take to get statistically accurate results.

Test Small Stuff – A lot of people think only things like the headline and call-to-action are important to test. But even the smallest things can lead to better conversions.

Test The Whole Thing – While sometimes a single element on a page can make a big difference in conversions. Making an entirely different version of a page can also make a huge difference. For example, once you’ve tested a couple of headlines and found one that converts. You can then test that headline in a completely different version of your page’s design.

Test Your Whole Funnel – A lot of people stop at testing the lead capture page or the sales page. But did you know that one lead page version might convert better, but result in fewer sales? It’s a good idea not only to test to see which version gets you the most opt-ins. But also which one results in the most sales later.

Send Quality Traffic – A lot of people think it’s okay to just send thousands of non targeted visitors to a page for split testing purposes. This isn’t going to work. You need to drive truly targeted traffic to your pages when split testing. Because any sales from un-targeted, low-quality traffic could be completely random. And have nothing to do with the elements you’re testing.

Real World Examples of Split Testing

Now that you’ve learned about the basics of split testing. Let's take a look at some specific examples in which split testing has yielded real world results for companies.

NuFACE is a skin care company that was getting good traffic. But people were abandoning their shopping carts at an alarming rate. They decided to do a split test to figure out if offering free shipping would decrease abandonment. The results were astonishing!

They found that by offering free shipping. They increased their orders by an incredible 90%, and also increased their average order value considerably. Those are some pretty amazing results!

Electronic Arts decided to test a promotional offer for a pre-order of their new version of the SimCity game. They originally had a huge banner at the top of the screen offering $20 off the customer’s next purchase when they pre-ordered the game. Assuming this would increase conversions, but what they found was exactly the opposite.

Their test revealed the variation with no promotional messaging actually drove 43.4% more sales! If they’d simply assumed the promotional offer would increase purchases. They would have lost a tremendous amount of money.

Express Watches (a UK-based watch retailer) added a “Seiko Authorized Dealer” to their product description pages. To help boost buyer confidence that they are buying authentic brand-name watches. This boosted their conversions up to 107%!

As you can see from these few examples alone, split testing has extraordinary potential to increase your bottom line!

Conclusion

No matter what results you’re looking for—sales, email opt-ins, etc.—split testing can help you maximize those results.

Please don’t think that split testing is something reserved for the “big dogs”.

Even if you only get a few hits per day. You can perform split testing. It may take a little longer to get the results you’re looking for, but it will still be worth it in the end!

To Your Success,

Joshua