Verfwinkel.nl is the largest online paint supplier in the Netherlands and has been working with Fightclub for years, including in the field of Conversion Rate Optimisation (CRO). Together, we have already carried out many experiments. Some were simple to build, others complex.
Verfwinkel.nl
Online marketing, conversion rate optimization (CRO), performance
One experiment was solidly based on data/literature/previous insights, the other just a fun idea we were keen to test. Great results alternated with setbacks and surprise. In the end, almost every test gave us valuable insights, such as:
No-brainers don't exist
Small adjustments sometimes have a big impact
Persevere, even if the first time is not a hit
Don't always try to change behaviour, but use it to your advantage
Wondering how we arrived at these insights? We'd like to explain this using three consecutive experiments around the Verfwinkel.nl shopping cart icon.
Insights
In a funnel analysis, using data from Google Analytics, heat maps and session recordings, our CRO team observed that a significant proportion of desktop visitors click on the shopping cart icon in the shopping cart and checkout of Verfwinkel.nl. When people do this in the shopping cart, the page reloads and they still see the shopping cart.
In checkout, clicking on the cart icon leads to a return to the cart. Not all visitors who return to the cart reach the checkout again. This results in an 11.5% conversion loss.
Why do so many visitors click on the cart icon, when it does not bring them one step closer to completing an order?
Or even a step further away from that order?
Possible cause: Zeigarnik effect?
This psychological theory states that unfinished tasks or unachieved goals often stress us out, making us eager to complete tasks as quickly as possible. The green ball in the shopping cart icon signifies an unfinished task for some visitors, think of the icons of unread notifications or messages on your phone. So click and complete as soon as possible!
This effect is very useful when the goal is to entice visitors to proceed to the shopping cart, but perhaps less suitable when people are already in the shopping cart or even already in the checkout. The question was how we could use the Zeigarnik effect to our advantage on most pages, without it being detrimental in the shopping cart and checkout.
A/B test #1: Less prominent icon
Our CRO team set about setting up an experiment, specifically an A/B test comparing the original with a variant. In the variant, the pink shopping cart icon and green ball were replaced with a less prominent green colour, which is also used in the button to continue shopping from the cart. It was expected that the shopping cart icon would be less noticeable, therefore less likely to feel like an unfinished task and lead to fewer clicks on the icon. As a result, visitors were expected to proceed to the checkout and complete an order more often.
Result & conclusion
After a four-week testing period, our CRO team was able to take stock. Result? No difference! Not in the flow from shopping cart to checkout and vice versa. Nor from both steps to a completed order. After a four-week testing period, our CRO team was able to evaluate based on sufficient data.
Result? No difference!
Not in the flow from shopping cart to checkout and vice versa. Nor from both steps to a completed order.
A/B-test #2: Hide icon
As with the previous test, it took little effort to build the variant. The expectation this time was that visitors were more likely to move on to the checkout and complete an order, as the ‘unfinished tasks’ disappeared altogether. We hid the shopping cart icon and entered another four-week testing period.
Result & conclusion
Another setback followed, even bigger than last time. We again saw no differences in throughput from cart to checkout and completed orders. There was also again no difference in traffic returning to the cart from checkout. Unfortunately, this variant from checkout did lead to 2.3% fewer orders than the original!
The rigorous idea was not a success. It seemed to ensure that visitors who wanted to make another adjustment to the product composition from the checkout could not return to the cart and therefore did not complete their order. Perhaps visitors used the icon more often than we thought on purpose anyway, with the intention of making another adjustment.
A/B-test #3: Redirect to checkout
We hadn't made any gains in two A/B tests where we tried to reduce the number of clicks on the shopping cart icon to improve throughput. What if reducing clicks was now no longer the starting point? And we could instead use those clicks to our advantage?
At one of the bigger players in Dutch e-commerce, a click on the shopping cart icon naturally leads to the shopping cart, unless you do it from the shopping cart itself... then it sends you straight to the checkout. Talk about using clicks to your advantage!
Our CRO team got back to work, building a third A/B test. Again simple to build and visually there was no visible difference, but in the variant of this experiment we sent visitors who clicked on the shopping cart icon from the shopping basket directly to the checkout.
Result & conclusion
After another four-week run, we finally had a winner! We found no difference in completed orders, but in the variant, 3.1% more visitors flowed through to the checkout. We had now used the clicks to our advantage.
Learnings
Besides a, ultimately, nice result, these three experiments led to some valuable insights.
No-brainers don't exist
We regularly get asked whether we should test certain optimisation proposals because ‘they're no-brainers, right!’? Some clients find our ideas so obvious that they ‘must be an improvement’. Or ‘if it works at a reputable e-commerce organisation, it must also work at ours.’ But that's not how it works. Despite our expertise and experience, even we are still regularly wrong. And making the shopping cart icon less conspicuous or hiding it may work for other websites, but not for Paint Shop.co.uk. That's why we like to test as much as possible, because these experiments show once again: no-brainers do not exist.Small adjustments sometimes have a big impact
Clearly, building these experiments was pretty straightforward. The differences seem marginal at first glance and sometimes there is not even a difference visually, but a small tweak in design usability can have quite an impact. After all, a few percent more or less does make quite a difference in large volumes on an annual basis.Hang in there, even if you don't get it right the first time
It's certainly not always a first-time hit, but almost every non-winning experiment yields new insights. When these are used in iteration to the next experiment, a winning result really does emerge. And it provides more and more insights into your visitors' behaviour.
Don't always try to change behaviour, but use it to your advantage
Initially, our CRO team focused on reducing the number of clicks on the shopping cart icon. Only after two experiments showed that this did not have the desired result, we changed directions. Visitors simply show a certain behaviour, whether you want them to or not, and this sometimes proves difficult to change. It may then be a better choice to adapt to this behaviour and make sure it leads to achieving your goal rather than trying to change the behaviour itself!
