Feeds and speeds vs. actual needs.
I am sure that there are trends to be extracted by staring at a bar graph or a spreadsheet filled with statistics. I think the truth is simpler than any CEO or Wall Street maven truly wants to admit: netbooks were the ultimate symbol that the PC has become commoditized. The Wintel platform and it's various hangers-on (Lintel, etc. etc.) have moved past just generic commoditization and into that dreaded next phase. That phase is the one where the item in question (computers, in this case,) is so ubiquitous that you now buy a device for a specific purpose rather than one generic device to solve all problems.
For a company like Dell, (or any major box pusher) this is an absolute disaster. It’s a return to having to compete with actual innovation. We no longer simply buy our computers based on what is the fastest, or even the cheapest. We buy the computer that suits our particular needs best. If you don’t find something in Dell’s line-up that meets your particular desires, you just shop around, confident that eventually you will find it. Like any good capitalist, Dell, Acer, HP, and every other box-shifter under the sun is fighting this for all they are worth. (ASUS, you and your eee-pc. Traitors! You caused such a ruckus.) Once they have to fight this battle on innovation, economies of scale are lost. Suddenly you don’t crank out 10M of some generic unit, offering a total of maybe 6 notebooks and 10 PCs for a given season. Now you have to broaden your range to offer more models, at fewer units produced per model. Not only that, but you need to offer variability in their specification, or you’ll lose out to some other competitor who decided they would try to make money giving people equipment they actually want to buy.
It’s a slow game, and big corporations like Dell, HP, Acer etc. are good at sending a tentacle to lash out at anyone who gets out of line. Still, it’s inevitable, as evidenced by, well, everything humans have made…ever.
You don’t buy a radio because it has the broadest range of channels, or the highest gain antenna, do you? No, you buy a radio based on it’s look, or number of storable channels, or maybe because it includes an iPod dock, or MP3 player or some other feature. I maintain that for the vast majority of users and of usage scenarios, we hit that point of "good enough" somewhere around the P-III 1Ghz. There will always be a call for "faster stronger smarter better" in certain niches, but I don’t need my e-book reader or remote desktop box to be stupid powerful. My HTPC needs to sit in the corner and not consume watts when I don’t want it to. My "morning newspaper" El-Reg reading laptop needs to only be capable of turning on, opening El Reg, and allowing me to read the articles, and post long boring comments.
Of the 15 or so computers I own, my home VDI setup, (sandboxing is good,) and my games rig have any oomph. (And the reality is the house server is configured for muchos big-time power saving anyways, since even hosting 4 VMs, it’s idle most of the day.) The rest of these systems could cheerfully meet every need I could possibly have for them with a P-III 1Ghz, or an Atom. I bought myself a new gaming rig in January. My next 4 planned systems, (including my replacement home VDI server,) are going to be Atoms. (Server gets a dual core.) I’ll be replacing 4 existing (but marginally flakey) systems with brand new gear that consumes a tenth the wattage of my previous systems for less than what it would have cost me to buy a new home server 5 years ago. Even for an avid consumer of technology like myself, I have reached the point where I spend more money every year on disks than I do on the systems that feed them. And I’m perfectly happy with the performance of all of that gear.
If you’re Michael Dell, that’s got to be ****ing terrifying.