A/B testing is a great way to test changes you want to make to increase your conversion rate. Below we are going to cover some of the most important points when it comes to validating an A/B test, as well as give you some advice for what changes you should make now.
Your approach to conversion rate optimization will depend on how much traffic you have. There are businesses with a lot of traffic and businesses with a little or no traffic.
Let's start with the businesses that have very little or no traffic.
If your business falls into this category where either you're just getting started out or you have little to no traffic coming to your site, you will not be able to gather many valuable insights from doing a traditional A/B test unless you have a large advertising budget and can significantly increase your website traffic.
Here’s why: When you run A/B tests with small amounts of traffic, they usually will not be able to generate enough sales for you to make a conclusive decision on which variant truly performed better. In short, there’s just not enough data to work with.
What should you do? Rather than taking a low traffic, low converting website and trying to A/B test your way to a better conversion rate one variable at a time, it’s best to just start off by implementing conversion rate optimization best practices on your website and then you can make changes from there.. Include a clear funnel for customers to engage in the checkout process, personalize the process, display social proof on your website, and have a lot of compelling product shots, etc
Use these best practices as your initial control and see if you can get your conversion rate up from there. A/B testing on a single variable while you're still small scale will take a very long time because there aren't a lot of visitors coming to your website so you won't get enough learnings to validate every single winner on your site.
How do you scale up from scratch? It’s a lot easier and a lot faster to just implement best practices and start from there. Then you can scale up your advertising and work on all of the other factors at play, such as the creative in your ad account, your email marketing backend, and your ability to drive traffic to your site to increase sales.
Once you get enough traffic to your site and once things are rolling a little bit then you can dive into A/B testing.
On the other hand, if you have a lot of traffic coming to your website. With tens of thousands or hundreds of thousands or even millions of visitors coming every month, you have a lot more flexibility to work with.
If this sounds like your business, install some kind of A/B testing software such as Google Optimize or VWO. (If you're on Shopify, there are a lot of great A/B testing Shopify apps.)
A good A/B testing technique for these large scale companies is to use a single variable.
Build a list of all of the things that you would like to test on your site. Do you want a headline change? How about the banner image on the home page? Have you tried testing different imagery on the homepage? Layout all the tests that you would like to execute over the next several months then rank them in order of highest impact to lowest impact. Start from the top and then work your way down.
Now the important thing here is to test one of these variables at a time. The reason being, if you’re doing multivariable testing, you are testing a ton of different things on your homepage all at once. It becomes more challenging to accurately analyze the data and assess which of those tests was actually responsible for moving the needle forward.
Just keep it simple. Test one variable at a time. Yes, it takes longer but if you zoom out and you look at three to five years down the road, you'll actually be able to increase your conversion rate faster because you used single variable tests, and you didn’t have to worry about muddied learnings and misinterpreting results.
If you're just starting out with CRO, don't be intimidated.
The important thing here is that you're moving quickly and using accurate data and interpreting that data the right way to drive your optimization efforts. CRO is something that's worth working on if you want to start generating more revenue from your e-commerce efforts. Remember, you don't need to completely overhaul your site before you start seeing results. So stay on top of your data, respond quickly, and have fun increasing your business's conversion rate!