Culture of  Testing

Culture of Testing

How A/B testing can make your company grow at rapid rate.

When we started our company our small team often had discussions about the design and functionality of our website. A green button converts better, a header with happy family will increase sales. Discussions were more than often won by those with highest rank or by the one eventually designing the portal, or our CEO, who we call the HIPPO ( Highest Paid Person in the Organization ) ;-). But since our company embraced Growth Hacking methods to achieve fast and huge Growth Results A/B testing has become a significant part in our decision making. In our Insights Blog, we share some of our experiment results to give you more insights in how we work, fresh from our Growth Hacking Team #1.

Experiment: showing less steps in our product funnel

The first experiment that we want to share with you, is an experiment on our funnel. But first of all, what is our funnel? Well, it’s our most important page. A page where our users orientate and customise their final product. A page with a lot of potential for page to page conversion rate optimisation.

How we started

We collect a lot of feedback and data before making any decisions. Via Google Analytics, our Customer Support panels, but mostly tools like Hotjar (heatmaps) and Fullstory to measure our customers' movement and behaviour. We found that our users (in particular new users) really get lost in our product funnel. It just has a lot to offer.. As a user you need to make at least five decisions in a row to drop anything in your basket. Pretty important stuff to be honest, but most of them really dropped off. Maybe it was just too much information. We decided on tackling this problem by setting up and experiment to hide everything but the active step. Our hypothesis was that users only focusing on the active step, were less distracted and the conversion rate to the basket page would increase with at least 5%. Variations for a/b test setup








Base (a) - open steps

Variation (b) - hidden steps

Experiment results

We have some rules for calling an experiment significant. It needs to run for at least two weeks with a minimum of 1,000 conversion per variation. We found out by hiding the steps the conversion increased with 16.69%. A big improvement and something we pushed live immediately.