Estimated read time: 7 minutes
A/B testing is a technique that marketers use to optimize their content and website experiences. Some people use the term split testing instead. Basically, the audience is split apart to see how people react to certain experiences. The A/B in A/B testing stands for the sample groups. One version (A) is served to one group of people. The second version (B) is served to a different group of users.
In theory, it’s an easy way to figure out what works best:
- Are more people clicking on the blue box or the orange box?
- Do more people read this article if I post the picture here or there?
- Does this call to action work better over here or over there?
Of course, I’m oversimplifying a bit here, but nonetheless, conversion-rate optimization matters. That’s what A/B testing measures and tries to improve: Are people taking the next step. Whatever the next step might be. Conversions can include:
- Form fills
- Email signups
- Clicking into a landing page
- Setting up a demo
- Purchasing a service
What is conversion-rate optimization?
In its simplest form, conversation-rate optimization means we update our web experiences so they make conversion easy for users. Chris Dayley, of smart-cro.com, shares insights on the topic with me on this episode of the Business Storytelling Podcast.
Make conversion easy and don’t overcomplicate things for the user. If there are too many options, people might not take any.
A/B testing tools: Google Optimize Demo
There are plenty of tools out there that allow you to A/B test your experiences. One is Google Optimize which you can use for free and is relatively easy to use.
- Go to the Optimize starting page
- Link your Google Analytics account
- Add the Optimize code to your website header
- Then follow the steps shared in the video.
The biggest thing to consider – as usual – before starting is:
- What is your goal with the test?
- What page might work?
- Can it be measured?
For example, as you can see in the video, for A/B testing in Optimize you can measure:
- An increase in pageviews to a next-step page. For example, this could be the thank you page after a purchase
- Against an event in Google Analytics
Thanks to Chris for this live demo on the topic.
A/B testing tools for other channels
A/B testing is also common in email marketing.
- Did more users with the B subject line open the email than the A subject line users?
- Who bought more? (If you are selling something.)
A/B testing should always relate to your business and digital goals. For example, if you are trying to sell a product, you could test with A/B testing which copy, design or other factor leads to more sales.
For example, some email newsletters work better – and are opened more – when they have very specific subject lines. But some audiences respond better to the same subject line each month. No matter how often you email, please respect the inbox and offer value!
Specific subject line:
Ten new social media tips to increase engagement
Same subject line every month:
Your monthly social media news from Christoph
A/B testing can help you optimize your content and digital experiences for your audiences and make the experience better for them and pay off for you in the long-run, too.
The future of A/B testing
A/B testing measures what technique works better for the masses. But here’s the thing when it comes to “the masses” online:
The masses, so to speak, will become less and less important and instant relevant one-on-one connections and experiences are becoming more and more important. Let’s break that down some more…
Most everything online in 2021 is a numbers game. The more relevant traffic you have, the higher your influence, reach, income, etc. etc. It’s still very much the typical marketing funnel approach – though not as linear as it once was perhaps. The more relevant people enter in the top, the more relevant people become advocates, customers, etc.
Certainly, there are exceptions out there, but in general that’s still the case. Think about Amazon, for example. They are hailed as a company that gets personalization right. And they do. They know me so well that it’s hard not to buy their recommendations. They are truly personalizing things to me. But, at the end of the day, Amazon is also winning the numbers game. Lots of people know about them and a ton buy things from the site. So, the importance of mass likely isn’t going away. It’s one of the most visited websites in the United States.
Moving to the on-site experience: Marketers measure what works better by doing A/B testing. When we do this, 2 percent more people do whatever we want to do. This is helpful and works in the pre-personalization area. In a web world where content is truly personalized A/B testing in the traditional sense won’t matter, because it’s not about moving the masses, but about moving the individual.
When two people have a relationship they know things about the other person – likes, dislikes, flaws, strengths, etc. Many of us – when we choose to – adjust our styles to the other person. the same concept – at some point – can apply online when it comes to digital experiences.
Instead of marketers testing what increases conversions from the masses, they work on knowing our customers and potential customers better. They try to have a meaningful relationship. What’s meaningful of course depends on each relationship and is a bit of a fluid concept. So that means that the content and experience presented to me is very different from you. They know so much about me that the experiences is completely optimized and super relevant to me. And by super relevant I don’t mean that it’s only optimized to get me to buy more. It’s optimized to help me – the customer. Of course, marketers knowing things about me, is also dangerous, because they may just use it to get me to buy, buy, buy. And sometimes it’s about buying something and sometimes it’s not.
Testing in the future
This is one of the more debated points I’ve made at conferences and in workshops. People tell me that this will never be possible and some have called me crazy. “Testing will never to away! Never. Never. Never.” I wonder if people whose jobs were automated in the last few decades said the same thing. And maybe in a perfect, futuristic world who knows if we’ll ever get there no testing is needed at all. Or maybe, as we are moving more and more toward personalization, testing moves from testing the masses to testing the personal experiences. So, some days I get served the A test and some days the B test. That could be very relevant, of course, to me, the consumer, if it’s used to make the experience more relevant to me.
Chris and I agreed that A/B testing will likely evolve in the far distance to something like this, but for now, we need and should test to see what is working and what’s not working.
Of course, A/B testing only works when you actually have people engaging with your digital platforms!