Implementing a Rigorous A/B Testing Plan for Ecommerce Success
Struggling to acquire new customers at scale? Learn more about how we've helped brands just like yours. Click here.
A/B Testing process.
In order to get the most out of your A/B test and also to ensure that you have a solid way of measuring and presenting success, its worthwhile getting a solid and repeatable testing process in place.
Here is a quick checklist that you can plan your A/B tests with.
Goal: Think about what problem it is you're trying to solve and base your goal around this. Example: it could be as simple as more conversions but could also be more website traffic from a certain demographic.
Hypothesis: What changes do you believe will help you achieve your goal? Example: Adding delivery information on the product pages will reduce basket abandonment rates
Length of test: How long do you need the test to run in order to have enough data to draw a concrete conclusion. This could be a length of time or a data point. Example: 4 weeks or until we reach 100 conversions
Variables: These are the elements that you are looking to chance in your B variant. Remember, unless you are testing and entire page variation, you'll want to control the amount of variables so you are able to pin point causes of success.
Metrics: What metrics will you be measuring? Choose a hero metric, closely related to your goal and use this as your north star. A group of secondary metrics can be useful but again, control the amount otherwise you'll have a sea of useless data.
Traffic split: Decide whether you want to split your traffic 50/50 between your A and B. Or whether you want to have a smaller test group versus your control group. Your goal will determine which will be more appropriate.
Once you have decided on all of the above, get started to putting your test together. We'd recommend avoiding testing too many things all at once as you'll want to be able to make definitive conclusions on how each element is affecting overall performance.
Ongoing testing process.
A/B testing like this is not a one and done solution. It should be something that is iterative and evolving over time.
Example:
We are testing 2 versions of checkout process and have determined during a 2 week test that our B variant where the form is split over several pages has higher performance success.
The checkout abandonment has dropped 7%. However we are still not converting 75% of checkout initiators.
We set the B variant form live and begin planning a second test on this same element.
Our goal is still increasing completed checkouts and our new hypothesis is that including our free delivery information at the beginning of the form will increase completions.
And so the testing process begins again. We can continue to iterate each element until we are satisfied that have reached the ceiling of success. This same process of testing and iterating can work across any element of your digital and optimising strategy.
A/B Testing Tools
There's lots of different online A/B testing tools catering to different experience levels and requirements.
Here are 5 we'd recommend checking out.
If you're interesting in learning more about data driven testing strategies, A/B testing or how to make the most out of your existing strategy, leave your details below and one of our experts will get in touch!
