This blog is missing a lot of details before we can consider this experiment successful. Most importantly, what is the usable throughput of a station in the "improved" coverage area? I would recommend using iperf before/after to see if there is any practical benefit.