Glad we agree on something! Actually, it applies to every mode; it's just that it's only one metric of many. But your point is well taken; instead of "spectral efficiency", I should have I said "the efficient use of the radio spectrum". Other important metrics include transmit power, where that power goes, how long I must transmit that power, and how wide an area must be cleared of other transmitters around my intended receiver to protect it from harmful interference. All expressed as "per unit of user data transferred". This is why "spread spectrum" can sometimes be more efficient than narrowband modulation, why error control coding (both FEC and ARQ) are so important, why automatic transmitter power control is seriously underappreciated...and yes, why source coding (aka data compression) is important. You're right, the proper measure is in fact the "goodput" -- how much user data actually gets through for a given amount of interference to other stations. I notice you didn't mention P4; while I am not familiar with every one of its modes it does seem to be designed to "shift gears" over a wide range to adapt to channel conditions. Many other modes, e.g. FT8, do not have any kind of gear-shifting ability and rely totally on a human operator to know when to give up and switch frequencies and modes. Adaptation to changing conditions is still one of those incompletely solved problems where ham radio can offer something to the state of the radio art. No, speed is important because you need to normalize everything to the total amount of user information being transferred. The JT modes are very well designed to operate over very poor channels, but their throughputs are severely limited even on excellent channels. The inability to take advantage of improved channel conditions, if only by decreasing transmitter power, is itself very inefficient use of the radio spectrum. Regarding channel sharing, you are referring to the "multiple access problem". A lot of progress has been made here, including by yours truly; an algorithm called MACA that I invented specifically for ham radio found its way into standard 802.11 WiFi. Spread spectrum can be used for multiple access; the technique is better known as Code Division Multiple Access or CDMA. It's one of the reasons why spread spectrum can actually be more efficient in practice.