Conversion rate optimization (CRO) is a system for increasing the percentage of visitors to a website that converts into customers, or more generally, takes any desired action on a webpage, according to Wikipedia.
So it is a countinous process that helps you to reach your goals.
To avoid the “I think we should . . .” sentences, which is usually just intuition. It can be OK in some way. Sometimes it generates more conversions, but the professional way is to make data-driven decisions. This is the main point for this post.
What is conversion
First of all, we should (briefly) describe what is a conversion. Conversion should be everything what is valuable for your business and every business has different conversions.
Of course, the main conversion for an eCommerce site is the purchase, but it has another conversions, too.
A registration, newsletter signup, info sheet download, contact form submit, etc should be a conversion at your business. Everything that has any value for the business is a conversion, so you should track these interactions. If you track them you would like to optimize, because there is always a room to optimize them.
Conversion Rate Optimization steps
There are several steps to reach your goals. To be honest these are not so difficult steps, but you have to go through every one to end up with the right result. So, please don’t skip one or two. If you do so it would be a guesswork, not a data-driven answer.
Measure
First things first. To know anything about your site, how the users interact with it, you have to set up tracking to measure the steps. It means you have to go beyond the basic Google Analytics setup. You should add enhanced ecommerce tracking and more in depth ones to know the details and avoid the guesswork.
If you track your site, but it doesn’t give you the answers, just go deeper and add more detailed tracking.
Analyze
If you have set up tracking, you measure the user interactions, so you have data. The next step is to analyze it. Try to find gaps, where the user left your site or too much data which can be some other anomaly.
If you have some data, but not too much, you can step backward and set up a more detailed tracking. Eg. you have a multi step checkout, you track every step submit, but realize that from step 2 to step 3 there is a huge drop more than the usual churn. In this point you should start to track every form element fill and add some heatmap and video recording about your users behaviour (do not record PII!)
You track the users’ interactions always, no need to switch on and off occasionally. It gives you the base. But heatmaps are rarely always on, you set up one as a campaign, to find why the users drop at step 3 on the checkout.
So, the first 2 step is “mixed” you have to back and forth until you have the detailed data, it’s like zooming in.
Create hypothesis
Now you have the right data, know the pain points, let’s create some hypothesis. Based on the data/info you have you can create a lot of hypotheses (options). Just write all of it down.
There is no good or bad hypothesis, write everyrthing donwn that come in your mind, the more is better.
It could be an endless game because there are a lot of things to change at an ecommerce site. If you focus on a micro/macro conversion, the endless game became a finite one because you have just dropped the majority of the hypotheses (on the next round you can focus on them :))
Some hypotheses:
- Change the order of the elements to get more focus the main form
- Make bigger/smaller a banner/text box/etc
- Make sticky the main CTA on mobile
Prioritize it
If you have more than one hypothesis (I bet you have more) you have to prioritize them by the impact of the change.
The first should be the one that could be the biggest impact on the site. I know it is a classic guesswork but you can test all the hypotheses with the help of the next steps, maybe the one you give the highest chance fails andthe second or the third will be the winner. It is OK, data tells you what is the best not the intuition.
Design it
If you would like to test a new layout/font/user flow you have to design it. It is not a big surprise 🙂
Ask your UI/UX team, involve your designer, tells them every detail and design the modifications.
Test it
If you all set, you “just” have to test the modifications. The test means you create variations and the test it agains the original one (and each other if you created more than one variation)
Repeat it
If you finished a test (and the whole process) you can and I mean you should start a new conversion rate optimization cycle.
There are a lot of improvement, you can made a lot of new hypotheses worth test. And, in a non-negligible way, the user behaviour changes all the time, what works today maybe lost it’s shine tomorrow.
Comments are closed.