The Challenge

Forever New, a leading Australian fast fashion brand, partnered with us to support their website redesign and replatforming, which included the development of a comprehensive conversion optimisation roadmap. Our partnership involved end-to-end user research and experimentation, all focused on increasing conversion rates and guiding key website redesign decisions. It was crucial that the new website not only maintained but also enhanced the brand’s online performance while addressing the diverse needs of all stakeholders involved in the overhaul.

Upon diving into the data and gathering user insights, our in-depth research for the brand uncovered several key findings:

  • The “Dresses” page, a high-level category page, attracted significant traffic from multiple channels, but its conversion rate as a landing page was low (GA4).
  • Usage of sub-category filters was low (GA4).
  • When users did interact with sub-category filters, the sales conversion rate increased (GA4).

Our Solutions

Based on these insights, we hypothesised that users were interested in shopping by subcategory, but the filter was either not prominent enough or difficult to use. By making subcategories more prominent on the high-level category page, we aimed to help users navigate the website more easily and find their desired products faster, thus increasing the conversion rate.

We proposed two experiments to test this hypothesis:

Experiment #1

Changes: A carousel with sub-category text links was added.

Result:

in sales conversion rate
+ 0 %
observed power
0 %
significance
0 %

Experiment #2

Changes: In addition to the text links, thumbnails were added.

Result:

in sales conversion rate
+ 0 %
significance
0 %
power = underpowered
0 %
Another finding of the study is the existence of “blind spots” in reporting. These result in economically relevant companies and entire sectors receiving only little media attention.
Tim McCoy Designer

Images did not help increase the sales conversion rate.

(Unfortunately, it’s not always feasible to just test a single change per experiment. Balancing speed and prioritisation of other idea is necessary)

Originally, the fashion retailer wanted to only test the version with text and images, following the best practice of highlighting options with images or thumbnails. However, after some discussion, we decided it was best to perform iterative testing to isolate cause-effects. Luckily we did just that, or we wouldn’t have uncovered such a great result. 

Key Takeaways

So why did the images perform worse?

It is theorised that the images were increasing the page load time. Another theory was that given the fast fashion nature of this retailer, the thumbnails used were often not a good representation of the sub-category. It may have been better to let the user narrow down to a subcategory first and just let them explore the styles.

All good learnings for future experiments!