Our Blog

Guide to Performing A/B Test in WordPress

AB testing in wordpress

Growth is a process of a trial and error.

Replace ‘growth’ with ‘Conversion Rate Optimization’ (CRO) and the quote still holds true.

CRO is the marketer term for UX. And no matter how many resources and case studies you’ve read online on Creating a better User Experience, if you’re any good at reading between the lines you’ll notice that nothing is universally applicable: What worked for someone else may (not) work for you.

That’s why the fool’s way is also the foolproof way of getting the most out of your WordPress website. That’s why A/B (aka Split) testing is a huge deal.

In this guide, we will take a deeper look at A/B testing and its application and implementation on WordPress websites of all verticals.

The Terminology:

A/B testing, or split testing, is a kind of live testing process where you pit two variations of a single page/post/title/interface element/theme/etc. against each other and measure how each of them performs.

Think of it as a hypothetical two-contestant version of American Idol. So who decides the winner from the two amazing singers? You leave that to public vote.

The Truths:

Unless your luck is astronomical, you’ll not hit upon the highest-converting design on your first try. You have to make a page, test it, and start tinkering until you get it right. A/B testing simply tells you what changes your visitors prefer (or what changes drives your visitors to convert).

Nothing more, nothing less…

Why you should be testing:

There’s always room for improvement: A/B testing helps you become better.

You’re targeting and converting your ideal users (or as close to ideal as is possible). Since you can’t be everything to everyone, you need to focus your CRO efforts on people who would actively engage with your brand instead of inflated numbers that mean nothing (lots of people know that those social counters on a WordPress website can be faked: like someone who Wannabe-cool).

You’re doing the absolute best with what you have: your website is doing what it was intended for. It’s functioning at optimum potential and therefore driving better results.

A/B testing can be an eye opener through the revealing insights: you will become more knowledgeable about your most profitable channels and your most loyal users.

Why Visitors Leave:

Visitors come and go through/for the following:

  • Outbound

Where are your visitors coming from is actually more significant to conversions than you might think. Outbound source-based segmented testing can be done through Google Universal Analytics and can be used for finding out detailed stats on your visitors from Social media, Search engines, paid advertising, promotional campaigns, referrals, and other sources.

Note: Your outbound sources are like the ‘public face’ of your business. It’s the context in which your visitor came by your site. Through testing, you can find out exactly which channels are profitable (for you) and focus more on those.

  • Inbound

Once your visitors have arrived at the site, what’s driving them away?

The problem could be anything: design, irrelevant (for visitor) content, a weird animation, the color, the parallax, etc… Every landing page/UX article will tell you to focus on improving performance, writing relevant/catchy content, amping up social numbers, ‘unclear’ call to action, etc…

These are usually the elements that most people start their testing with.

The Process:

Keep in mind that testing should be based on clear facts and not whims of designer and/or marketing teams. Here’s how you do it:

1. Define a goal: A realistic, practically achievable goal. This will help you decide which elements to test.

2. Hypothesis: What – The elements you’re testing. Why – Are you testing those elements. And what are you hoping to achieve (the intended outcome). All these points should all be backed by data.

For example: Testing – Post Headlines; Because – They’re the first things people see (and if you could throw in some feedback, you score triple); Intended Outcome – Better engagement.

3. Make a thorough check of your time and resources.

4. Set up control group and a variation.

5. Run the test

6. Review results

7. Implement and repeat

Having an ‘ideal customer’ persona can really help you with targeting and quality leads/conversions. Tools like Helpful Canvas can be used to do so.

Understand demographics, countries, age and jobs, device preferences, active surfing time, etc… of people who are already using/visiting your site regularly too.

Also remember to prioritize, because even though there are infinite number of things to test (more on that below) does not mean that you have the resources and time to do so. Test things with higher conversion value first.

What to Test:

Anything and everything can be tested, as long as you have good reason (valid data, legit customer feedback, etc.) for doing so. Here are some good elements to start with:

  • Headlines
  • Featured Images
  • Landing Pages (and all its own elements: copy, images, videos, placement, social buttons, etc.)
  • Menu
  • Sidebars
  • Headers and Footers
  • Call to Action buttons
  • Typography
  • Mobile layout
  • Content (length, tone, formatting, etc.)
  • Design (theme, colors, CSS, etc.)
  • Ads and their placement

You can scrounge the internet for more tips and points.

Testing tools:

Here are some A/B testing tools, made with love for WordPress Website Development:

Title Experiments:  A simple to use plugin with a conversational, jargon-free tone; something which will endear it to novice testers immediately.

Optimizely: A platform that lets you “test, personalize, and optimize web and mobile apps”. It has a plugin that integrates with WordPress, and it is very feature rich.

Nelio A/B testing: Just as (if not more) feature-packed as Optimizely, also lets you test WooCommerce product pages.

AB Press Optimizer: Useful suite of necessary features, clean and efficient to a ‘T’.

Cohorts.js: An amazing JavaScript framework for A/B as well as multivariate testing. Great for enterprising developers testing JS interfaces.

Google Analytics: Because nothing can beat this. Lets you run any experiment you could dream of, if you have set it up properly with your WordPress website.

Endnote

Every experiment you run will have one of two outcomes: NO difference; or SOME difference (further divided into positive or negative). The key is to keep testing and implementing the positive results until you’re somewhat satisfied with the result. And then start over.

Repetition is a form of insanity, but in this case, it’s just how you grow. Who cares? We’re all mad here.

Leave a Reply