Why reverse-engineering the marketing moves of successful companies is a bad idea…
Let’s talk about A/B testing.
A/B testing is the name of the game if you are involved in digital marketing in any way. But only a few know that this now popular tool saw its origins in the 1920s. This was the time when the statistician and biologist Ronald Fisher first introduced this experimentation method to the researcher community.
We have come a long way since then of course. Now if you type in A/B testing to Google, 3 out of the 5 top organic results are all about digital marketing (and one of the other two is Wikipedia…)
We will listen to Google and stay in the most popular domain, digital A/B testing. Hotjar defines A/B testing as “the act of running two different versions of the same website as part of a controlled experiment and gather data on which performs or converts the best.”
There is a myriad of different areas we could touch on here but we want to keep this post a digestible size. For now, we collected the rules that you do not want to break when running your A/B tests.
🟡 Have a clear & well-founded hypothesis
A/B testing is not for generating new ideas. You want to start out with an evidence-based and well-research hypothesis with clear and different options.
⚪ Define a clear baseline
You need to establish a baseline against which you can later compare the performance of your test. You should look into the performance of the version live on your website. This very well might be the winner no matter you are only testing a single potential update or multiple.
🔴 Avoid biased experiment design
It is easy to fall into the trap of designing the whole experiment with a biased mindset. Never start a test with the end goal of proving what you think will work.
🔵 Be clear on the goal of the experiment
Is your goal to increase the conversion rate? Or maybe you want more traffic to a particular page? Both are valid and testable options. You just need to be clear on which one is true for you and measure the success of your A/B test with the relevant metric.
⚫ Collect data from enough users
The more data you have, the better. You always want to aim for your test to be statistically significant to draw valid conclusions.
All in all, A/B testing is a method for giving you more data. And you need data to make educated and informed decisions. You never want to just go ahead and copy what your competitors or successful brands are doing. Chances are they are not serving your ideal audience.
You need to test first and then implement it.
If you are interested in how A/B testing can be used in your SEO campaigns, click on the link below.
(We are running our beta currently so message me for 3 months free access)
