I wanted to provide a follow back on the banner issue.
Joseph Seddon has been _incredibly helpful_ on this. Whilst he confirmed
he indeed drop the banner visibility to its minimum on purpose (sigh),
he took time to propose and implement a new banner proposition and a AB
testing system to test banner efficiency.
Lot's of appreciation for his help on this. I apologized to him for what
was over-reacting. It was the disappointment from discovering what was
wrong after several days wondering.
For those with loads of curiousity
Results for the AB testing may be seen here :
With the new banner, Desktop see's a 100-150% improvement in traffic.
Mobile is a mixed result. With an slight 5% improvement in English but a
2-5% decline in french.
The new banner :
The meta page :