IBM used a 5-team product design process in the '60s onward, according to one exposÃ© IBM-insider book I'd read. Five teams developed The Next Generation (of whichever generation ahead was studied) and Management would take the Middle Option. Not the Best, not the 2nd Best, but the middle ground won the day. The latter two were tossed BUT the hype machinery would say, "Our next-gen is so much better! We thought about THIS or THAT but no, here's our better solution!" and would point to these 4th and 5th-place finishers.
IBM claimed they had an immediate replacement with Second-Best and Best choices, too.
The question was, however, "Why not always select The Best?"
Answer: "We have no faith in our engineers, or engineers of the future. What if today's Best is THE best and we can't ever develop a better one? We have no faith in our employees or the brains of the future engineers."
Of course, no one in an IBM boardroom would say that out-loud. Not for the record. But in practice, it was the Act, not the words that displayed this.
I have wondered if IBM would have been a bigger player in today's consumer computer world if they'd have used The Best in each generation, and then beat the war-drums louder to seek true improvements, each and every time.
Or did this "We have no faith in our employees" take a toll and those engineers started fudging their designs. "Why knock myself out?" was indeed a quote used from IBM engineers.
Take, for example, the era when Steve Jobs left Apple and suddenly the PowerMac and PowerPC were delivered. The PowerComputing Company had a huge future - the merger of Apple's few engineering resources, plus Motorola's heftier section, plus IBM's massive section. A huge future. "Unix is the OS of the future!" was every other sentence coming out of their mouths.
Then Steve Jobs returned and strangled them, killing off PowerComputing Inc and such a bright future.
AMD was a smaller x386 clone-maker at the time, and suddenly was able to hire all the newly laid-off PowerComp staff that couldn't find homes back at M'Rola or IBM. Apple soon shut down their CPU development side.
And then there were only two... and now, only one is talking about a viable CPU future.
(Of course, that's the Desktop World. IBM still does tons of mainframe CPU work - or what they call mainframe - but I hope this is merely a tidal episode and, when it comes back in, the tide will put up many more processor-development companies - like Via, C, Samsung, M'Rola still - albeit in the handheld or set-top-device world.)