The drawer case — a usability analysis on Contentsquare
Measure before you rebuild: validating a design decision with quantitative data
During peak season, a drop in conversion cast suspicion on a central component: the drawer, a sliding panel inherited from the design system, used to present the mobile homes.
Rather than redesign it on a hunch, I set up a multi-segment quantitative analysis on Contentsquare with a Product Owner, to measure how it was actually used before passing judgement.
The data cleared it: exposure, scroll depth and click-through were all excellent. The friction was elsewhere — and the investigation could restart on the right track.
Outcome — The data protected a structural design choice instead of sacrificing it on an assumption — and redirected the investigation toward the funnel's real friction.
When conversion drops, the reflex is to look for a culprit. And a visible, central, slightly ageing component makes an ideal one.
Behavioural analysis ended up clearing a component we were ready to rebuild. That’s what a data culture is for first of all: not breaking what works.
01 · Design
A component born from a content constraint
The drawer — a sliding panel — has been part of the design system since its inception. When the time came to design how mobile homes were presented, it imposed itself for three concrete reasons.
First, it let us make thin content count: the database was sparse and photos were rare. The drawer’s long-form text format turned that constraint into useful information, where an image gallery would have rung hollow.
Second, it stayed simple to maintain for the content-production teams who feed these pages day to day.
Finally, by keeping the list of mobile homes visible in the background, it signalled to users that they could move quickly back and forth between descriptions without losing track of their search.
02 · Doubt
A suspicion, and a refusal to rule on intuition
During peak season, a significant drop-off appeared between the product page and add-to-cart. The question asked itself: what if this drawer, so obvious on paper, was in fact misunderstood, misused — or outright blocking?
The easy option was to rebuild it. Instead, I set up a multi-segment quantitative analysis on Contentsquare with a Product Owner, to test the suspicion against real usage. Four segmentation axes, chosen to separate the component’s effect from that of purchase intent:
- stay dates entered or not;
- mobile home configured or not;
- new visitors or returning ones;
- mobile or desktop.
03 · Measure
The data pushes back
The analysis flatly contradicted our fears. Scroll rates and time-on-component were particularly high: the drawer was seen, opened, read through. Click-through on the add-to-cart CTA turned out to be strong too, confirming genuine interest in the different mobile-home layouts.
Above all, the segmentation revealed a sharp distinction in purchase intent: 40% progression to cart among users who had entered their dates, against 12% without dates. Two different behaviours — committed buying on one side, exploratory browsing on the other — not a usability defect.
The conclusion was unambiguous: the drawer’s usability could not explain the drop-off in the funnel. The friction was elsewhere.
What this demonstrates — when conversion falls, it is easy to blame a central component and feel like you’re acting. Here, the data protected a structural design choice, and spared us a costly rebuild that would have fixed nothing.
04 · Reopen
What I’m looking for now
Clearing the drawer doesn’t close the case, it moves it. I recommended two directions.
First, measure the discovery rate: what share of users actually scrolls down to the mobile-home section at all? An excellent component nobody reaches is still a problem, just a journey problem rather than a component one.
Second, run moderated usability tests on the rest of the product page, to find the real friction points in the booking journey.
I could have rebuilt the drawer and ticked the “we took action” box. Measuring first spared me a pointless redesign and told me where the real friction was.