Internet marketers commonly utilize split testing to measure the effectiveness of a marketing offer.



It's a scientific method that allows you to exactly determine what variations of individual elements of a web page can increase its conversion ratio (i.e. the percentage of visitors converted into sales).

Different split testing methods

1) A/B Split Test

It's the simplest one. Basically visitors are randomly directed to one of two different versions of a web page and results (sales, leads or click-throughs) are tracked.

This method will tell you what page is better, but not what individual elements on the page have positively influenced its conversion ratio.

You can run several split tests in a row and test different variations, but this can be a very time consuming process.

2) Multivariate Split Test

With this method you can test multiple combinations of different page elements at the same time.

Let's say you want to test the effectiveness of two different headlines (h1 and h2), product descriptions (d1 and d2) and product images (p1 and p2).

Your website traffic is evenly split so that each visitor will see one of the possible page versions made by a particular combination of these different elements.

Visitor A sees h1-d1-p1 and doesn't buy.

Visitor B sees h2-d1-p2 and doesn't buy.

Visitor C sees h2-d2-p2 and buys!

And so on for every possible combination. Once the test has run for enough time to give statistically significative results, you will find the winning combination of those page elements that produce the optimal conversion ratio.

In the example above 8 different web pages need to be tested.

Let's say you want also test two different testimonials (t1-t2) in combination with all the other page elements.

The pages that need to be tested are now 16. You can easily see how this number will exponentially grow as you add more page elements to the test.

More are the pages to test more traffic you need to generate enough conversions to determine the ?winning? page.

This is the main drawback of multivariate split testing.

Page elements that can be "split tested"

To limit the number of page variations and reduce the amount of traffic needed to generate significative results, split testing should be used to test only those elements on a web page that are most likely to have a large impact on the conversion ratio.

Headlines. You can test different versions of you main headline or unique selling proposition.

Bulleted lists of benefits. You can test if and how sales are affected by different product's benefits or by listing benefits in a different order.

Product descriptions. Test if and how different product descriptions, each focused on a different product benefits affect sales.

Product images. Split testing will help you determine how different product images, their number and quality influence sales.

Increases of up to 400% in conversion ratios have been reported by marketers utilizing split testing methods. Obviously split testing won't tell you how to write a good headline or an effective web copy, but will help you scientifically test all your marketing ideas and put them together in the best possible combination.