How To Conversion Optimize And Avoid A/B-Testing Failure
What is the right way to improve your conversion rate? Sleeknote decided to dig into that question and had a chat with Michael Aagaard, who is Senior Conversion Optimizer at Unbounce.
Michael has more than 7 years of CRO experience, which means he has made a lot of mistakes – and learned from them.
He believes that CRO is much more than just split testing, and in this interview, he shares his strategy for how to conversion optimize your business in the best, most efficient and powerful way possible.
Conversion optimization is a broad term, and it includes many different perspectives. I sat down for a chat with Michael Aagaard, who sees conversion rate optimization (CRO) from an entirely different angle and has developed a method of successfully optimizing countless businesses.
Michael Aagaard is the Senior Conversion Optimizer at Unbounce in Vancouver, which is a software service tech company that has developed a landing page builder that helps marketers build customizable and powerful campaign landing pages without IT.
What is conversion optimization?
There are many different interpretations, and most have their own idea of what conversion optimization really is. According to Michael, CRO is not just split testing – and they are not one and the same. You do not have to do split testing to conversion optimize. Michael therefore says we need a better term for conversion rate optimization.
The conversion rate is just a metric we use, but it is not the only metric. Our goal is to optimize our businesses. CRO is the practice of understanding users and creating a conversion process. Research is what helps us understand our users, and that is where Michael spends 80% of his time.
The process of conversion optimization
During the process of conversion optimization, e-commercers tend to make huge mistakes when trying to conversion optimize; get excited and start making random changes to the website.
But that is not the proper way to go about it, according to Michael. Understanding your problems before you try to solve them is crucial. You should spend more time identifying your problems rather than solving them.
When Michael gets a new client, he analyzes the site’s data, walks through an analytics review, sees how much traffic there is on the site, and determines how many conversions they have. In order words, he investigates the basics to gain a better understanding of the case.
He then moves on to figure out where traffic gets clogged up and where there are leaks. This gives him an idea of where and what the problems are. Only then will he move on to why – understanding why these problems occur. He does that through customer interviews wherein he asks for people’s expectations, concerns, and worries.
Once the research is done, he’ll find a potential solution and test it out. By testing, Michael doesn’t mean that you should split test like crazy. It’s more about carefully measuring the effect.
Optimize everywhere
When you want to optimize your business, and you start doing your research to find and solve your problems, you must think of all the possible ways in which you could make your customers convert more often. That includes offline optimization. Listen to Michael’s opinion of offline optimization and how you can do it:
Michael’s learnings
Michael has more than eight years of CRO experience, which means he’s made a lot of mistakes – and learned from them. Here are a few of Michael’s most important learnings that are worth considering when conversion optimizing.
First of all, do your homework – conversion research is fundamental in the optimization process.
At first, Michael thought he was a scientist because he was running tests, but all his tests were based on guesswork. You should only test based on thorough research, so when you fail, you’ll learn something, because you know the underlying reasons as to why you tried that one thing.
Don’t do split testing without understanding the statistics. Split testing is statistics!
Where many ecommercers fail
Do you ever wonder why your tests aren’t as effective as you thought, or why you don’t get any useful results? That’s probably because you need to do more research. Michael said that most e-commercers just guess what might work and split test it.
Also, there are millions of blog posts with easy ways or fast ways to optimize your e-commerce sites. They are no good because you don’t know the statistics behind the arguments. Here is Michael explaining what you should avoid doing:
How long should you run a test?
To find out how long your test should run, you must ask yourself, ‘When is the result significant?’. This is where we go back to statistics. As Michael said, it’s complicated, but worth learning.
You have to know what p-value, power level, and many other terms are, and you must know what they mean.
The more traffic you have, the quicker you can get a representative result. Do the calculations up front. There are plenty of test duration calculators out there you can use.
These are the things you must know how to calculate the test duration:
- The baseline conversion rate on the page you’re testing
- Have an idea of the minimum lift you’re looking for
- Confidence level
- Traffic per day
There is also one other thing you need to consider. Hear what it is here:
Tools to use
Here are Michael’s favorite tools:
Analytics – having access to analytics and data are vital as you can’t do CRO without data!
Hotjar – A cool suite of conversion tools all-in-one.
Hotjar’s Feedback Polls – Michael’s number one go-to tool. This is a fantastic way of getting insight.
Interviews – Yes. Real, personal interviews with customers, sales people, support people, etc. Make this a permanent part of your research process.
Wrap up
Michael points out that the tendency (and the most common mistake made) within CRO is that people forget the basics and move on to the next big thing. This leaves a lot of half successes.
He hopes that people will go back to the drawing board, as the core competencies should be part of the future of CRO. People should step away from the automation aspect, as it is still a people-to-people industry, and understanding your users is important.
This article originally appeared on Sleeknote.com