Conversion optimization isn’t just about optimizing your website and increasing your conversion rate. Even though it might be the end goal, there is so much more that goes into the conversion optimization process.
It’s easy to feel overwhelmed by the amount of data and testing you have to go through to find and solve problems. Thus, I had a chat with Karl Gilis, who’s an expert on usability, and he has an entirely new perspective on the conversion optimization process that could potentially change your conversion optimization strategy.
Karl is the co-founder of AGConsult, a Belgian usability and conversion optimization agency, and he has been in the usability business for more than 15 years. Karl is all about facts and proof rather than opinions and trends, which is surely one of the reasons that he has been voted as one of the most influential experts within the CRO. He made his first website in 1996, and it was about everything that had to do with computers—he was quite a nerd at the time.
Here are the major takeaways from the Karl Gilis interview:
- Start looking at 50 of your user session recordings when you want to define your problem area. By looking at user session recordings, you can immediately see how people behave and where there might be a problem.
- Test, test, test! Even though you think you might have found the perfect solution, there’s always a chance that you can find an even better and higher performing solution.
- Don’t copy tests. Your website is different from your competitors’, and just because a test worked well for them, it doesn’t mean you’ll get the same results on your site.
Want to know a funny little story about Karl?
Download this interesting story about Karl Gilis and learn what his first conversion optimization epiphany was.
Don’t be a copycat
A lot of conversion optimizers are betting against best practices and things you read online, but Karl thinks that it’s still a good point of inspiration. You shouldn’t copy a test, but you’re allowed to be inspired by it.
You get your ideas on what to test online, but it’s not the place to start. For Karl, it’s more important to ask what you should test instead of looking for ideas. Listen to Karl guide you through the first steps of testing:
What to look for and where to find it
Google analytics is always a good place to have a first glance. As Karl puts it, he’s not a data scientist, but he usually looks to see if there are any differences in conversion rates or visitor behavior between different browsers and devices.
Second, you should check your drop-off rates. Determine the places where people leave your website, and check to see if there are any differences in bounce rates.
It’s not about whether the bounce rate is high or low or if your website is good or bad; it’s just a starting point to get an overview of your website’s performance.
It’s much more important to look at user sessions. It might be more intensive to look at, but it will reveal a lot more about how people behave on your website.
For most people, user sessions are an easier way to see the problems. You have to be very good at reading data and analyzing data to use Google analytics in an effective way.
By looking at user session recordings, you can immediately see how people behave. Are they using the search in the right way? Are they typing the wrong words? Is there autocorrect? Is the autocorrect suggesting different words than expected? What are the search results? Is the order of the results logical? And so on.
User session recordings will reveal a lot more than Analytics will.
Of course, Google Analytics can be a very effective tool, but it has also become more and more advanced and difficult to use. Thus, for the average shop owner who doesn’t have a dedicated analytics team, user session recordings are the best starting point.
Karl explains that Analytics is good for the obvious things, and user session recordings are better for the less obvious things.
For example, if you have a funnel in Analytics and you can see that only 2% leave the funnel from step one to step two, and 78% leave the funnel in the next step, then you can clearly see that there is a problem between step two and three.
Back to basics
A lot of e-commercers don’t look at the differences between browsers and devices.
During the late 1990s, there was a browser war between Netscape and Internet Explorer, which, later on, became Internet Explorer 6 and 7 and, then, Chrome. Some web agencies decided that they wouldn’t support Internet Explorer 6, which resulted in a loss of 20% of the users, who were still using Internet Explorer.
If you compare it to a physical shop, it would be the same as closing the door to 1 in 5 people and not letting them into your shop. If you have 1 million visitors and don’t give access to 5%, then you lose that 5% of your income.
Craig Sullivan made an important point at a conference that Karl had recently attended, which was: What is the use of adding persuasion techniques if the visitor can’t use your website?
Start with the basics: make sure that your website works for everybody that everyone can find what they’re looking for, and, then move on to persuasion and personalization. The end result of personalization on a shitty website is a personalized shitty experience.
Are you really ready for A/B testing?
One of the problems that many e-commercers face is that AB-testing has been hyped so much in the past 2 to 3 years that everyone wants to do it, but only a few websites and e-commercers are truly ready for it.
A/B testing is about refinements and adding smaller bits and pieces to make your site work even better.
You first need to use user testing and expert reviews to cut out all the unnecessary elements on your site, and then you can begin to test. As Karl so finely put it: First, you cut the crap, then you can A/B test.
Valuable user session learnings
In 50% of cases, Karl will start with user session recordings after he thinks he’s noticed something in analytics. For example, if you have a page with a higher drop-off than expected, something is probably wrong. Listen to what Karl does in the other 50% of cases:
Another valuable lesson to be learned from Karl is that it can be useful to delete all the sessions that last less than 10 or 15 seconds. These are usually people who enter your site and leave it immediately again and thus have no value to you.
You found a problem – what’s next?
That’s the difficult part because now it’s more about experiences than hardcore facts. Once you’ve found a problem, but you still don’t know the underlying reason, you have two possibilities according to Karl:
When you finally get to the testing, there are three outcomes:
- The test doesn’t reveal anything
- The test is a failure
- The test is a big win
If your test doesn’t reveal anything, do something completely different and test again. If your test still doesn’t show anything, it’s probably because the element that you’re testing isn’t that important to your conversion rate, and you shouldn’t waste time on it.
Karl suggests that you simply remove that element completely because it doesn’t have any purpose or effect.
If the result of your test is negative, you have still learned something important. The element that you’ve changed clearly has an effect on your conversion rate, and it’s now up to you to find another idea and test that.
If it’s a big win, that’s great, but you should still do another test as your next test might reveal an even better solution. You can never test too much.
Those are Karl’s three steps with A/B testing. Remember that every time you do a test, you learn something new and gain experience.
Don’t get fooled by the competition
It is easy for companies to say that they increased their conversion rate by 70% with just one test, and as a result, many will copy that test. What people fail to recognize is that the company has presumably done 3-4 tests in between that were huge failures in order to get that conversion rate.
Your website design, your audience, your products, etc., are all different, and using the same test will never get you the same results. Listen to one of Karl’s major learnings with A/B testing:
Where on earth do you start with mobile optimization?
One of the biggest problems with mobile optimization is that when you convert to a new platform or device, it’s easier to take the existing website and transfer to the new in a smaller version. However that can create a lot of problems with usability and user experience.
It’s imperative that you look at how people behave on a mobile screen. Most people use one hand, or even just their thumb, to navigate through websites on their phone. Thus, you have to consider where you place the most important elements of your site because they have to be in reach within that one finger.
Some might find it easier to reach the bottom of the screen with their thumb, while others might find it easier to reach the top of the screen. There’s only one way to find out what works for your site – test it!
Karl’s Favorite Tools
For user testing, Karl uses one tool, which is himself. The other tools he uses are:
- Hotjar: It’s a good entry point because it has user session recordings, scroll heatmaps, click heatmaps, and you can do targeted service all in one tool for a very reasonable price.
- Google Analytics: Collects all your data in one place, so you can quickly get an overview of how your website is performing.
- Heap Analytics: Once a user has given their email address or logged in, you can follow that user all along, and it is very easy to set up.
Remember that the tools are just tools. They won’t provide you with a solution. You’re the one who has to connect the dots and find a solution through logical thinking. Thus the most important tool is your brain!
What is the future of conversion optimization?
First of all, four years ago, everything was about A/B testing. Now, it’s more about optimizing the website. What many fail to recognize is that it shouldn’t all be about the website and conversion optimization.
You also need to focus on your product range. If your products aren’t selling, you should think about making changes there instead of just focusing on website optimization.
Conversion optimization needs to be more embedded in the company culture. It should have an influence on product range, product offerings, etc.
Secondly, the tools we use today will become much more intelligent, and artificial intelligence will come to play a much bigger role.
The basic tests will become completely automated within the next five years. For example, what Karl refers to as the layout tests—the design and color of your website—will be completely automated.
Furthermore, conversion optimization will become more specialized in product development and product range. However, influential techniques such as psychology, psychology principles, copywriting, persuasion techniques, etc. will still be handled by humans and not computers.
That was all from Karl Gilis. Conversion optimization isn’t just about making your website look better; it’s also about doing the necessary legwork to understand and define your problem areas. Remember that just because you find a good solution that has a high conversion rate doesn’t mean you should stop testing to find an even better solution. Test, test, test!
Karl surely provided us with a new angle to conversion optimization, and hopefully, you’ve learned a trick or two.
What tests have you done that failed miserably? And, what have you learned from them? Share your story in the comments below, and educate your fellow marketers.