Want to Sell More Online? How 'Split Testing' Your Site Can Help Here are four simple tests that compare different elements of your site to help it generate more leads and sales.

Your website is a critical tool to help your startup to attract new customers and sell more products and services. But if your site design isn't optimized to engage visitors, you might be missing out on leads, sales or newsletter signups -- otherwise referred to as conversions -- without realizing it.

One way to improve your site's performance is through "split testing," a strategy where you test two versions of website elements to see which works better. There is more to split testing than simply comparing two versions of a page. You need to measure the results and see if there's a statistically significant difference between the two variations.

Let's say, for example, you ran a split test on a landing page and came up with these results:

• Variation 1: This version received 903 visits, with five conversions
• Variation 2: The modified version of the page received 976 visits, with seven conversions

At first glance, it looks as if variation 2 is the better page, but don't jump to conclusions. Split testing is based on measurement, and your results need to be statistically significant, meaning the results you've achieved are due to measurable differences not random chance. Otherwise, it can be harder to prove that the difference in performance will continue.

Related: 3 Ways to Create More Engaging Website Landing Pages

For split testing to be successful, you need to have a large number of eyeballs on your pages, as you'll need a certain quantity of visitors to track whether or not your target conversions are happening. If you don't yet have enough traffic to produce statistically significant results, focus on a traffic growth strategy first and hold off on split testing until you have more visitors. But if you have more than 10,000 visits a month, get started with any of the following split tests:

1. Headlines.
A headline can significantly affect the number of conversions your page achieves. Just consider the experience of Chicago-based software company 37signals, which relied on split testing to increase signups for Highrise, its customer relationship management software.

Its original headline said, "Start a Highrise Account." Centered below it was the line: "30-day free trial period on all accounts." Highrise split test several landing pages and eventually settled on the one with the headline, "30-day Free Trial on All Accounts." The subhead read, "Sign-up takes less than 60 seconds. Pick a plan and get started."

The revised headline brought a 30 percent improvement on conversions, the company said in a blog post.

When you test headlines, think about how you can benefit the prospect. Headlines that are clear and enticing generally are better for increasing conversion rates.

2. Social media calls to action.
Split testing also can help you craft more effective messages on your website to attract social media followers. For instance, Dustin Curtis, founder of San Francisco-based publishing network Svbtle, decided to test different messages to try to get more Twitter followers.

First, he tested, "I'm on Twitter," which had a 4.7 percent click-through rate. But he settled on his second message, "You should follow me on Twitter here," which produced a 12.8 percent click-through rate. If Curtis hadn't elected to test this particular call to action, he would have lost potential followers by sticking with a less effective message.

Related: The Psychology of Color and Branding (Infographic)

3. Images.
You can split test different types of images on your site. If you put pictures of real people on your site, for example, it's likely that visitors will be more engaged. Studies have shown that website visitors react more strongly to pictures of faces than to any other type of picture.

You can also test different colors on your site, including the colors of the buttons visitors click. Toronto-based healthcare company Carelogger increased its conversion rate by 34 percent, and all it did was change the color of its sign-up button from green to red.

4. Checkout pages.
Testing doesn't stop on the home page or product pages. It should be happening on all the pages you deem important -- from landing pages to checkout pages.

For example, Calgary-based e-commerce site Hockey Canada Store wanted to increase conversion rates and reduce the number of people who were abandoning items in their online shopping carts, so it decided to test its checkout page. By switching from the original two-page checkout to a single page-checkout, it saw a 21.8 percent increase in conversion rate -- likely due to the decreased complexity of the checkout process.

These are just a few of the numerous split tests you can try. Don't be afraid to try something new because those are often the tests that can yield a home run. And remember to always keep an open mind. If something isn't working, don't be afraid to test it again.

Related: 5 Things That Belong on the Front Page of Your Website