If you are investing time or money in search engine optimization (SEO), online advertising, or any other method to drive traffic to your website, then you want to make sure your website is designed to convert visitors to customers as quickly as possible.

The goal of any online marketing campaign is NOT to generate visitors.  The goal is to generate customers.  So the key data point to measure and improve is your website conversion rate.

In this article I’ll explain the steps to improve your conversion rate using webpage split tests.


What is a Split Test?

Let’s quickly review what is a split test for anyone not familiar with this technique.

A split test is the process of splitting website traffic between two or more different versions of a webpage.

For example, if I had a 2-page split test (also known as an A/B split test) running on this page you’re reading, then 50% of my visitors would see one version of this page and 50% would see a different version of this page.

Since the website traffic is randomly divided among two different versions of the same page, we can measure the difference in some of the desired actions we want on the page to get statistically significant results.

In my example of this blog article, I could measure the time on the page to determine which version is more engaging to my readers.

Over time I would eventually collect a significant sample size to determine with confidence that one webpage design outperforms the other.

Now that you know what split testing is all about, let’s talk about how to use it to boost website profitability.


1. Define the Webpage Goal

The first step is very important because you can’t improve anything unless you know what the goal is.

The overall goal of the website is to generate more sales, but each webpage has its own objective to push a visitor through the sales funnel.

So you have to look at each webpage individually and define the action you want a visitor to take that will lead her further down the sales funnel.

For example, the goal of a “contact page” is to persuade the visitor to call or complete a contact form.  The goal of a “services page” may be to first complete a form to access a free guide or whitepaper that pre-sells and pre-qualifies the visitor.

As you can see, different pages will have different goals and it’s important to define the individual goals before you start any split test.  We’ll use these goals to measure the conversion rate of each page.

The conversion rate is simply the number of people who completed the goal divided by the total number of visitors.  If your goal is phone calls and you received 5 calls from 100 visitors, then your conversion rate is 5%.


2. Identify the Webpage Elements to Test

Once we have our goals defined, then we need to identify the webpage elements we’re going to test.  This is where you need to be creative, but at the same time be willing to try something you personally may not find aesthetically appealing.

It doesn’t matter which design you like best or what your gut tells you is going to work.  The beauty of a split test is your market will vote with their actions to tell you exactly which design is best for your business.

So don’t be afraid to take a risk.  Sometimes the design you hate will turn out to be much more profitable than your pretty version.

I recommend you test everything, but here is a list of very important elements:

  1. headline and sub-headline
  2. prominent and clear call to action (test wording and location)
  3. location and size of webforms (test number of form fields)
  4. images

If you’re not sure how to improve your webpage then ask a friend who is not familiar with your business to take a look and tell you if something is not 100% clear.

Or if you have no friends — just kidding ;) — then check out Usertesting.com where they will video live people trying to use your website.  This is invaluable feedback to help you improve your website.


3. Install Split Testing Software

The next step is unfortunately a little technical, but there are a lot of programs available that make this pretty painless.

When you’re just starting out I recommend you use Google’s Website Optimizer (GWO), which is free and easy to set up using their “setup wizard.”

Simply follow the steps in the wizard and then select the option to send the code to your webmaster if you’re not comfortable editing HTML.  Once your webmaster adds the code and you verify the different pages are functioning properly, then you can activate the test.

Another tool I would recommend is Visual Website Optimizer (VWO), which is free to try for 30 days, but then you’ll need to pay at least $49/mo to keep using it.  VWO is much easier to set up because they have a nice WYSIWYG editor and you don’t need to generate new code for every test.


4. Get Significant Data

The final step once you have the test running is to sit back and wait for significant data.  Both GWO and VWO will tell you when one of your test pages has outperformed the other page by a statistically significant amount.  At that point you would stop the test and push 100% of your traffic to the best webpage.

However, some tests will drag on forever if both pages perform about the same.  In this case, I recommend you wait until you get at least 100 conversions on each version of the webpage.

If there’s no clear winner after each version has generated 100 conversions, then I recommend you start a new test by changing different elements on the page.


Get Started Now

If you’re not currently testing to improve your website, then get started today.  This is the one tactic you can use to improve sales without increasing the number of visitors to your website.

If you have any questions, then post them below.