John Cox writes about why enterprises may not want increased bandwidth on their WLANs: Maybe I'm missing something, but I don't really see why increased bandwidth will add tons of complexity for network managers. Seems to me that network managers who may want to add voice in the future would be happy for the added bandwidth, which should increase the number of users in range of a given AP.
My basic point is that adding more bandwidth in the current WLAN environment isn't as simple as adding more bandwidth in a wired LAN.
Shifting from 11b to 11g shifts you from about 5-6M bits/sec throughput to 17-22M bits/sec, so the users associated with a given access point have more total throughput, which is then shared among the users. But the limiting factor in some cases is the the number of available channels, not the throughput: 11b and 11g have only 3 non-overlapping; 11a has 9 or 11.
More bandwidth/throughput *may* be needed if you add voice. But again, as with most things in WLANs, it all depends -- it's complicated. Bandwidth may be less important than quality of service -- being able to give priority to voice traffic. And the IEEE group working on QoS is...still working on QoS.
I didn't mean to imply that more WLAN bandwidth is Evil. I recall the adage that "bandwith covers a multitude of sins." But that doesn't seem be quite as true for WLANs.