A/B Split run testing

Eliminating the guesswork from your website…

• How Split Testing Works
• A/B Split Testing
• Multi-level split testing
• Which option is best for you?
• Taguchi Multivariate testing
• Designing the experiment

How Split testing works

Split run testing is a marketing technique for comparing the effectiveness of multiple variations of a web page. In a split test, your visitors are ’split’ between two or more variations of a web page - seeing either version ‘a’ or version ‘b’. Real time results provide immediate feedback on which variation is more successful.

Split testing traditionally has been used by direct marketers for years. It is an exciting new tool for the internet - eliminating much of the guesswork on what does and doesn�t work.

A/B testing:

A/B testing allows you to test two landing pages, or email newsletters, against each other in real-time.

Multi-Level Split Testing:

Allows you to test only the things on your page that are most likely to have a large impact on results.

Multi-level split testing is like A/B testing on steroids. A multi-level test lets you focus on only the things most likely to have a large impact on effectiveness.

The multi-level test focuses on one variable on your page (eg a heading, an image or special feature) - offering several different options for each. In testing terms, this is known as testing one “factor” at several “levels.”

A good headline can easily get you a 50%+ gain in conversion, meaning a test on your headline is often a good place to start.

Variable rate split testing allows you to significantly reduce or even eliminate the financial risks associated with testing.

Variable rate split testing allows you to split your traffic at any percentage you would like.

Which option is best for you?

Consider the following split test scenarios:

Scenario 1.

You have a landing page that is working well but you want to improve it:
In this case, your current landing page is your ‘control’ offer. Testing at 90%/10% effectively minimizes your risk, while allowing you to test out new ideas in an attempt to improve your results.

Scenario 2.

This is a new offer, you don’t have anything to risk:
With a new offer, there is often nothing to risk. Running an AB split here will work fine, but as soon as your test begins to show a winner, consider modifying your percentages to give preference to the better performing variation.

Taguchi Multivariate Testing:

Imagine a tool that could show you how to build the ââ?¬Å?perfect salesââ?¬? page. That tool exists today, and its known as Multivariate testing using the Taguchi method. Through sophisticated mathematical formulas, Websalad will show you not only what works and what doesnââ?¬â?¢t, but also why things work the way they do, and how to build the best page possible.

Find out WHY a certain page works…
Traditional testing helps us determine the effects that subtle changes have on response rates. Unfortunately, the only way to conclusively know what caused improvements in response is by testing one change at a time. This can be time consuming and costly over the long run - particularly if you have many ‘factors’ to test.

Multivariate testing goes far beyond this. In one test, you�ll be able to test many variables at the same time, and learn what changes have the most impact.

Designing the experiment:

Websalad will help you design a test by providing a structured process which asks you to enter the variables you would like to test and the possible values for each. Our system then designs your experiment and provides html recipe templates to be uploaded to your site.

Each template defines a specific way to build a landing page.

Each landing page recipe is then tested, and the results are tabulated in real time. When the test has concluded, Websalad will give you the “golden recipe,” the combination of variables that will give you the absolute highest possible conversion rate.

Taguchi testing is the equivalent of running thousands of split tests simultaneously. It can increase your sales response by 300%-500%.
PH (02) 8080 7475 or to determine the best option for you.