The Roadmap To Successful Conversion Optimisation
Conversion optimisation is often approached incorrectly… In this article, I’ll show you how not to do it. But more importantly, how it must be done. Step by step.
Do you recognise yourself? You read an article in which Company X increased its conversion rate by 50%, by changing the colour of the call-to-action button from red to orange. You happen to have red buttons on your site as well. So you set up an A/B test to see whether orange buttons would work better for you. In the worst case, you don’t even do an A/B test and you just make all your buttons orange in a hurry. Let the money roll in!
But what happens? Nothing. Absolutely nothing. No difference at all.
And so you decide: CRO doesn’t work! Wrong. CRO works ridiculously well. Only not that way.
How should you do it?
Here’s where the example went wrong. The majority of conversion optimisation efforts are not based on research. And thus not based on data. A data-driven approach to conversion optimisation guarantees you the best results.
Often a conversion optimisation attempt starts with 1 of the 2 following scenarios. And both scenarios are doomed to fail:
- You read a case study and try to repeat the success from the case study to your site by testing exactly the same thing. That’s like going to the doctor with knee pain and asking for a certain painkiller because that helped the neighbour with his back pain. Every site is different. Different problems. Different solutions. You can only prescribe a remedy if you have made the correct diagnosis.
- A so-called expert tells you immediately what you need to test. Yes, he has experience. And he will probably hit the nail on the head more often than someone without experience. But that’s like a doctor looking at your knee from a distance and deciding that you need a cast. The doctor might be right. But the chance is also great that he has prescribed the wrong remedy. He can only reach the correct diagnosis with some X-rays and other investigation. And then prescribe the proper remedy.
Without good research, you’re flying blind. You just do something. And that is a waste of time, money and traffic. You can better use your traffic for A/B tests that do have a chance of succeeding. Only in that way can you make it as clear as possible what works best for your visitors. And only in that way can you increase your turnover.
But how does such conversion research go exactly? Follow these 3 steps:
a. Correct data collection.
Make sure that you collect the data in the right way. Most likely, you have Google Analytics (GA) on your site. But is GA configured in the right way? Are your goals set correctly? The goal-directed steps? Your filters? Is event tracking set where needed? And so forth. Do a complete ‘health check’ of your Google Analytics. Because if you bring in data wrong, you’ll draw conclusions on the basis of bad data. And that can have disastrous consequences.
b. Quantitative research
Is your data coming into GA correctly? Super! Then we can now dive into GA to see where it’s going wrong. A good conversion optimiser will need 1 to 2 days to draw ample information from your GA. Your optimiser will play with segments, reports, devices and browsers until finding out where things go wrong.
GA of course never tells the complete story. That’s why with quantitative research you also use mouse tracking tools such as SessionCam or Clicktale. With those, you make (among other things) heat maps, scroll maps and click maps that give you a clearer view of exactly what your visitors are doing on your site.
c. Qualitative research
Here is the focus of your research. The quantitative research will tell you WHERE things are going wrong. The qualitative research will tell you WHY.
Do the following:
- Send a customer survey to recent customers. Try in that in particular to find out why they bought from you, or why not. Send the survey only to new customers who recently made a purchase for the first time. From them, you get the most valuable feedback: the purchase was recent, so they still remember the process. And in addition, they haven’t gotten used to using the site (like your loyal customers have) and thus you can better learn what the stumbling blocks were.
Don’t waste any time and money on expensive tools for this. With a simple Google Forms document, I have often achieved fantastic results.
- Do user testing. You will get very valuable feedback from it and can see live where your testers struggle, what they misunderstand or do wrong.
Talk with the customer support division. Or if you have live chat on your site: go through all the transcripts from the live chat.
Add a web survey to your site. Qualaroo is one of the best tools for this.
- Cross-browser testing. Test your own site in different browsers, on different machines and different operating systems. Test it completely. Yes, in principle your web agency should have already done that. But we all know that this is not always done thoroughly. Cross-browser testing takes some time. But it can be very useful. If you note that your checkout, for example, doesn’t work on IE9, while 20% of your visitors still use IE9, then it might be clear that you could earn a lot of extra conversions by handling this problem. A handy tool for this is Crossbrowsertesting.com, with which you can simulate all possible device-browser-operating system combinations.
- Appeal to an expert for a heuristic analysis. A heuristic analysis is an analysis with which an expert makes use of frameworks and methodologies to evaluate your site. This will more quickly identify the potential flaws of your site. Take note: everything that is found in a heuristic analysis must in principle be supported after the fact with data from the rest of the qualitative or quantitative research or confirmed (or ruled out) on the basis of A/B testing. Never limit yourself to just a heuristic analysis.
- Do a usability analysis, or appeal to an expert to do this for you.
2. Customer theory
Now that you have collected all the data and gone through it, you begin to get a clear picture of who your customer really is. Not per se who you thought the customer was. But who the customer really is. Write down your customer theory. This is a description of your customer. You can do this in the form of one or more buyer personas. How detailed you are in this is up to you. As long as your customer theory is a handy working document that you can refer to later. With each decision that you make, you must keep your customer in mind, after all. And you must thus test your decision with your customer theory.
Your customer theory is not static information, but a dynamic document. With each split test that you carry out, you will learn more about your customer. And you add that to your customer theory. In that way, you can for example learn from a test that your customer is not sensitive to discounts, but is sensitive to free delivery. You add that insight to your customer theory. In that way, you get an ever-more-complete picture of your customer. If you understand your customer better, you will serve your customer better. And better service means earning more.
3. Action plan
The next thing you do is to set up an action plan for all the problems that you have discovered from the research. Some things are so obvious that you don’t have to test them, but can implement them immediately. Other things will not be completely clear and demand further research. And there will of course be a number of problems for which you have a hypothesis about why it’s going wrong. You’re going to test these hypotheses.
4. Test, learn, convert.
Only now is it finally time to start testing. In contrast to many others who have had no success with conversion optimisation, you set up your test only when you have a clear hypothesis. A hypothesis that is based on your research. You no longer test randomly in the hope of finding a winner. No, you set up tests with a much higher chance of success because they stem from a better understanding of your customer.
Always start with an A/B test. Only consider multivariate testing if you really have a lot of traffic (more than 100,000 visitors per month).
Granted, not every test will produce a clear winner, certainly not in the beginning. But that is not that bad in and of itself. As long as you learn something from the test. And in that way, you further expand your customer theory. The more clear your customer theory becomes, the better your next tests will be and the higher the chance of success for the next tests. As strange as it sounds: the ultimate goal of your test is to understand your customer better.
Because understanding your customer better = serving your customer better = earning more.
Tools that you can use for this include Visual Website Optimizer and Optimizely, the 2 most-used AB test tools.
5. Keep testing
Don’t stick with 1 version of a test. Even if there is a clear winner, who can say that it can’t be even better? Put the winner in a new test and compare it with another version that stems from the same hypothesis. From your first test, you learned something about your customer, for example that the customer responds better to free delivery than to a discount. The version with the free delivery is thus the winner. But what if you test that version again, against another version in which free delivery is advertised even more prominently? Maybe that will produce a few extra percent.
Keep testing constantly. You can always learn something more about your customer. And you can always further improve your conversion rate. Maybe you’ll get lucky and double your conversion rate with one or a few tests. But the chance is greater that you will need many tests, each of which delivers a few percent improvement. When you add them all up, you eventually get to a drastic increase in your conversion rate.
Hopefully, this article has given you some structure and guidance for your conversion optimisation efforts. Conversion optimisation is not simple, but if you follow this roadmap, you’ll go a long way. And remember: ‘There are no losers in the testing world. If you’re learning, you’re winning’. Good luck!