Different Elements to Split Test on Your Website

-

 

As we’ve mentioned before, split testing is a powerful way to improve the conversions on your website, ultimately resulting in more sales and more money. If you’re interested in taking advantage of this technique, read on for more information on what split testing is, what elements on your site can be split tested and how to get started with this procedure.

What is Split Testing?

Split testing refers to the process of serving up different variations of site pages or site elements and tracking the conversions that result specifically from each option in order to improve the overall effectiveness of the site.

If that all sounds complicated, don’t worry – it’s much easier to understand when you take a look at an example. Say that you’re thinking about running a promotion on your site and want to know whether people will respond better to a headline that says “20% off all items” or “Free shipping on all purchases.” Instead of just guessing which option will convert better, you can install a split testing script which will randomly display either option to your visitors and track which headline results in more sales.

Once you’ve determined which headline is most effective for your readers, you can replace the losing headline with yet another variation to continually improve your conversion rates.

This process is referred to as A/B split testing, and it’s the easiest type of split testing to implement on your site. Simply select two variables, set up a test using a split testing script like Google’s free Website Optimizer tool, and you’re off. Or, if you want to test multiple variables at once – say, the color and text of a “Subscribe” button on your site – you can use multivariate testing, although this process is a little more complicated and is better suited to more advanced website owners.

What Should You Split Test?

Now that you know how to use split testing to improve the effectiveness of your site, let’s look at some of the different elements that you can test on your site. Sure, headline and button text are two pretty obvious examples, but there are actually plenty of other elements that can be run through this process.

Background color – Although most webmasters select their site colors based on personal preference, color theory is actually a powerful force acting on your site visitors. Because different colors can provoke different subconscious emotional responses, your site could be sending the wrong message to your visitors without you knowing it. For this reason, consider a split test that pits one background color against another.

Page headline – Since the average web visitor only spends about seven seconds deciding whether or not to read more of your content or click on the back button, making improvements to your headline is one of the most powerful things you can do to increase your site’s performance.

Type of offer – Whether you’re selling physical products or trying to convince people to opt-in to your mailing list in exchange for a free gift, you’ll find that visitors in different niches will respond better to different types of promotions. For example, sites selling weight loss or skin care products often do better promoting a “free trial offer” than they do giving away a free ebook. Similarly, visitors in the gaming industry who are tired of being charged for shipping with every purchase might respond better to a “free shipping” offer than a “20% off” coupon – of course, you won’t know until you test!

Offer text – The specific words you use to present your offer can have a major impact on how successful it is. As an example, consider the following two checkout button text options from ViperChill.com. Simply switching the offer text to “Get Cloud Blogging: Just $37!” from “Limited Time Offer, Just $37” resulted in a nearly 4% improvement in conversion rates.

Button text – 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.

Web form content – If you use a web form to solicit subscribers for your email newsletter, you’ll want to test the different elements on your form, including the headline text and the number of data entry fields you require. You may find that changing the call to action of your web form header or the number of fields subscribers are required to enter could dramatically alter the number of subscribers you’re able to capture.

Web form placement – In addition to testing the content that’s included on your web form, you should also test how your response rates changed when the form’s location on your site is changed. Consulting heat maps to determine which areas of your site are looked at the most may also help you to determine the best possible placement for your web form.

Images – Pictures can be powerful motivators in terms of capturing a person’s attention or provoking certain subconscious thought processes. And the best way to figure out if your pictures are serving their intended purpose is – you guessed it – with split testing. While this isn’t the most important element to test in terms of the potential for change in your conversion rates, it’s still a good one to look at if your other changes haven’t led to the response rate you’re looking for.

Keep in mind that this is only a partial list of the different elements that can be split-tested on your website. Just about anything can be tested – from the font you choose to the color of your headline text to the overall width of your page – although the specific testable elements that are likely to result in the biggest changes to your conversion rate are your headline, your offer and your button. Start with split tests on these elements and then work your way towards improving more obscure items.

How Do I Get Started with Split Testing?

Really, the only thing you need to set up these tests on your own site is a script that allows different versions of your page to be served up at random. Google’s Website Optimizer (as mentioned above) is one program that does this for free, although there are plenty of other options as well – ranging from scripts you install on your hosting account to split testing programs that offer real-time tracking and analytics.

Whichever program you choose, the most important thing to keep in mind is that you’ll want to test a single variable at a time (unless you’re using the more advanced multivariate testing). If you change too many elements at once – for example, if you use an A/B testing script to serve up two versions of a page with completely different headlines, offers and buttons – you won’t know which specific elements led to higher conversion rates.

So instead, to get started with split testing, pick one variable to focus on and wait until you have enough data to determine which version is the statistically significant winner. Once you’ve determined this, remove the losing option and replace it with another version to test. Over time, the information you generate by honing in on the most effective version of you page will enable you to dramatically improve your conversion rates.

Image: glogster

4 Responses

  1. Joel

    Great Article! I’ve found background testing actually helps a lot, and in over 20 tests I haven’t found a single background that beats plain white on a landing page. We’ve tried everything from really fancy patterns, images, even professionally designed themes that cost $1000+ and still none have out performed a white background.

    That being said, it’s always good to test it just in case, and in some markets it may be completely different :)

    We use Zentester for testing which I’ve found is so much easier to use than GWO and can be used for free too.

    Also you can test multiple things at once, you just have to use Multivariate testing rather than a/b testing.

  2. Ossie Gorin

    Here are more jobs…Mobile advertising reps wanted to support the Think Local Shop Local movement.Must have either an iPhone, iPod Touch, iPad or Android OS smart phone. Good pay flex hours no startup costs.sales.adxplore.netHigh School and college students welcome.

  3. How to SEO Your Life @sujanpatel

    [...] an example of how this can play out in real life, I used A/B split testing to determine that getting 6-7 hours of sleep a night works best for me.  Your results may be [...]

  4. The #3 Most Overlooked SEO Skill @sujanpatel

    [...] made plenty of mentions both here and on the Single Grain website about how important the process of A/B split testing is to a website’s success.  But the thing [...]

Leave a Reply