Landing Page Split Testing

No comments

Split TestingIn the previous two posts, we examined when you need to use a landing page, and how to create an effective one. Whether you’re using it for a PPC campaign, lead capture, or a featured promotion, you have a particular goal in mind, and everything on the page should be designed to accomplish it.

Everything – the images, the colours, the headline, the main body text – should actively move your visitors to do what you’d like them to do. Typically, that means either clicking through to the next page, or submitting their contact details.

If you’ve utilized the proven strategies mentioned in the previous post, you’re well on your way. Your chances of a successful campaign are good. So you just sit back and wait for the customers to start rolling in, right? Wrong.

Once you’ve created your landing page and it’s “live”, you’re still not done. There’s an axiom about online marketing: Always be testing. Any page you make could always be better. More efficient. A tiny tweak here and there could lead to significantly higher click-through and conversion rates, and that can amount to much higher sales or leads captured. How do you find the best version of any given page? Through landing page split testing (also called a/b, a/b/c or multivariate testing).

What is A/B Testing?

A/B testing refers to a comparison of two or more slightly different landing pages. You have your “A” version (called the control) , and your “B” version (the variation). The basic idea is to use both pages for a set period of time, and then compare their results. It’s crucial that you have an active goal for the test – something that people actually need to do – such as visitors submitting their email address. That way, you have data to measure and compare. You can determine which version is performing better (more people doing what you want them to do). “More visitors” is not a useful goal, as there are too many variables that could affect the outcome, and the differences on each page don’t have anything to do with it (after all, they’ve already landed on the page before they actually see it).


How Do You Conduct an A/B Test?

There are several methods you could use. Some of them are free, and some have a monthly fee (but do offer a more robust and user-friendly experience). Visual Website Optimizer and Optimizely are two of the better known paid services. They allow you to quickly and easily set up, conduct, and analyze tests with little or no knowledge of coding or analytics.

If you want to go the free route, there are many services and plugins available for your WordPress site. Install, and then follow their individual instructions.

A third option is Google Analytics. It provides a detailed analysis of your page traffic, sources, and conversion rates. This method demands at least a general understanding of the service (a topic for another day), although you can use the Google Content Experiments plugin to make things a little easier on you. You can set up basic A/B tests under Behavior-Experiments on the Analytics main page.

No matter which method you opt for, an A/B test follows the same general structure.


Here are five steps you need to complete:

1. Ask a Question
2. Formulate a Goal
3. Design the A/B pages
4. Collect Data (for at least one week)
5. Analyze the Results

A good understanding of landing pages will help you ask the right questions. It’s best to keep things simple, at least for your first few tests. You might, for example, ask whether different short copy on your call-to-action (the text on the submit button) would lead to higher conversions. Or whether an orange button placed above the fold (on the top half of a web page) would yield a higher click-through rate. Or whether a different headline would increase sales.


There’s no limit to the variations you could test:

● Headlines (text itself, one vs two lines)
● Sub-headings (no sub-heading, different sub-heading)
● CTA Text (text itself, on-button vs below button)
● CTA Button (colour, shape, size, placement)
● Images (number, placement, different images vs no image)
● Testimonials
● Social Media (button, icons, placement, displaying share and like numbers)
● Design Elements (colour schemes, framing, amount of whitespace, order of text)
● And much more…

The key is to ask a doable question and designate a measurable goal. Formulate it as an actual question in the early stages. “Would a blue colour scheme with a large orange button yield higher conversions than what we are seeing now?”. Then, create the “B” page, changing only the colour scheme to blue and the button to orange. It’s important to leave everything else exactly the same, so as to know that whatever change you see – either positive or negative – was a result of that variation. Limit your variations to one or two (if they’re small) at the most for each test. Higher than that, and it’s impossible to determine what exactly had the effect.

A typical A/B test will then split the traffic, usually 50/50 (with half of all visitors seeing the control page, and the other half seeing the variation), during the length of the test.


How To Analyze the Data

At the end of the test period, you’ll have enough data to accurately compare the two pages. You’ll want to look at the total number of unique visitors, the conversion rate (how many of those clicked-through to the next page, or submitted contact details, or made a purchase, or whatever other goal you set), and how those numbers compare. If, for example, your control page (“A”) had a conversion rate of 13%, and the variation page (“B”) clocked in at 17%, then you have a 4% increase. You would be wise to make the variation page your new control. It’s performing better. That’s it. You’ve conducted your first split test and acted on the results. A negative conversion lift means you don’t change the page. Simple.


Why Is A/B Testing Important?

It’s a numbers game. Even though the actual difference is only 4% in the above example, that represents a 26% improvement. Would you knowingly turn away 26% more sales or leads?

Look at an even smaller change. Let’s say a control page resulted in a 2% conversion rate, while the variation page saw a modest increase to 2.5% overall. That’s not bad. But what is 0.5% in the grand scheme of things? Again, you’ll want to think in terms of conversion lift, and going from 2% to only 2.5% is an improvement of 25%! And all because you changed the image appearing on the page (for example). Are you starting to see the benefits now?

Smart marketers know to always be testing. Always tweaking. Always experimenting. You’d be surprised at the little things than can result in big, big gains. You want to see your conversion rate going up. You want to capture more leads. You want to see higher click-throughs and sales. A/B testing can help optimize your landing page for maximum results. Find something that gives you a boost, and then ask another question. And another. Always. Be. Testing.


There are plenty of services and tools for conducting A/B tests yourself with relatively little headache or tech knowledge. If you’re still feeling uncertain, contact us and let’s chat about your marketing and how we can help.

ExhaleMarketingLanding Page Split Testing