Experiment: Room Type Carousel

Setup

In this experiment we introduced a filter and carousel for the different rooms (living room, dining room, bedroom, etc.). We tested the hypotheses, whether more user-friendly facets (the room for the new carpet) and a visual presentation (carousel) improves interactions, (micro)conversions and overall retention.

Desktop

The experiment had three arms:

1. noCarousel: base with no filter and carousel
2. filterOnly: only filter enabled
3. both: both filter and carousel are enabled

Feature Usage

• when enabled, this facet the most used one (although it’s pretty much on the top, so not so surprising)
• but with filterOnly option, it’s never selected in the filter panel (-> need to understand why):

When looking at the individual aspects, people selected the first two (livingRoom and entranceHall):

Experiment results

• trending: more returning users, more viewer page
• micro-conversions about the same
• engagement & retention trending positive; conversions still low

Overall, the experiment confirms our hypothesis:

• having filters that better match the user needs is beneficial
• the carousel presentation itself shows mild improvement in retention