What Are The Most Important Items To Split-Test On Your Website?

When it comes to conversion rate optimization, the work really never ends. There are always things to test and things that could be improved. Knowing where to begin or on which areas to focus, especially if time, resources, or money are issues, can be a challenge, however. Plus, once companies find something that works, they often neglect to test the particulars to see if their conversion rates could be improved any more. And this is a big no-no.

Never settle—even if you think what you have is working well.

That said, here are the top items on your website that you should split-test in order to improve your conversion rates:

  • Unique selling proposition (USP)
  • Landing page text length and position
  • The checkout process
  • Button elements
  • Pricing models

Not All USPs Were Created Equal

While all the selling points for your products and services are valid and have their merit, they are not all weighed the same by your customers. Some hit close to home, triggering an emotional reaction and a want to buy, while others elicit nothing more than a “meh” response.


Image Source

“Test all USPs in your ads, prominent placements of the site and social media to figure out which of these improve your conversion rate the most,” says Rocco Baldassarre of 1DollarAd.com. Then, once you discover which USPs perform the best, you can create an advertising and communication strategy that will allow you to maximize your level of customer attainment.

Pay Attention To The Length And Position Of Your Text

In theory, you would think shorter copy would fair better than lengthier prose in just about every contest. However, as with many aspects of the ecommerce world, there is no one-size-fits-all solution, and that’s why you need to be diligent about split-testing.

For example, “After testing his personal website, Neil Patel found that long-form copy produced 7.6% more leads (better-quality ones as well). On the other side of the spectrum, a Scandinavian gym chain got 11% more conversions with shorter copy,” reports Stephen Walsh of Lightship Digital.


Image Source

The position of your text can also affect conversion rates, so you need to think about where you’re going to place your testimonials, call-to-action, and information about the features and benefits of your products and services. Rocco Baldassarre suggests the following: “Use metrics such as sales and average time on site to evaluate the performance of your pages.”

Furthermore, don’t cling to “golden rule” guidelines, such as “your should place your most important information above the fold,” because for every rule-of-thumb, there’s a contradicting case study, like the one from Content Verve, as reported on by Kissmetrics, in which placing the CTA below the fold resulted in 304% more conversions than the control treatment.

Image Source

Streamline Your Checkout Process

Customers should be able to get in, find what they’re looking for, and check out with minimal effort or hoops to jump through. The parts of the checkout process that you absolutely want to test are guest checkouts, billing information forms, security seals, and single-step checkout pages.

How Persuasive Are Your Buttons

Here’s what you should be testing with your buttons:

  • Color
  • Size and shape
  • Placement
  • And of course…

eric siuEric Sui of Single Grain“The specific wording of the buttons on your site can also have a major impact on how well your offer performs. As our case study on the marketing lessons learned from the Obama campaign demonstrated, experimenting with different button text options (in this case, things like ‘Donate Now’ vs ‘Please Donate’ vs ‘Contribute’) resulted in dramatic improvements in donation rates, so don’t just assume that the default button text is automatically best.”

Pricing Models Go Beyond Dollars And Cents

Aside from how much you are going to charge for your products or services, you need to consider pricing models as well, like freemiums, free trials, and money-back guarantees. And with the latter two, you also need to think about what length of time is going to work best—7 days, 14, days, 30 days, 90 days? You need to test!
free trials

Image Source

How do you feel about split-testing? What other items do you test on your website?