The Perfomas -- oh God, the Performas
After Apple disgorged Steve Jobs and brought in ex-soda CEO John Sculley, the latter got the idea to spew out many SKUs of Macs. This was the Perfoma line, designed to be less intimidating than the Mac itself (intimidating?), but the sheer landslide of barely distinguishable models was intimidating enough. With the same basic hardware, there were educational models, direct sales models, models for sale at a mass-market retailer ... each software bundle might be a little different from the others, but who could keep track?
Also, it didn't help that most Perfomas were, well, crap. The quality ranged from not so great to awful. For example, the 4400 -- a "fat pizza box" desktop -- was supposed to be targeted at casual business users, but it was so poorly built that peripherals would suddenly not work, hardware glitches would cause hard crashes, and so on.
Needless to say, this adversely affected Apple's image of providing high-quality products. Soon after his second coming, Jobs made quality Job 1, or something like that. He also quickly stripped down the product matrix: one consumer laptop, one pro laptop, one consumer desktop, one pro desktop. There is no Step 5.
The cloning vats
From 1995 to 1998, Apple tried something new: licensing. It's been an article of faith among the "Apple will die ... any day now" crowd that the company made a fatal mistake in restricting Mac operating system use to actual Macs. The idea was that if Apple became just an operating system vendor, like Microsoft, it would grow like topsy, like Microsoft.
Shaky logic there, but Apple tried it. Starting in the non-Jobs era, I should add.
The strategy did have some salutary effects, not the least of which was enabling Power Computing's awesome ad campaigns . Daystar Digital experimented with then-unusual multiprocessor configurations; some lower-priced Mac clones hit the market; some companies pushed the build-to-order and direct sales models; and Power Computing armed itself with ex-Apple engineers to push a few technical boundaries.
But the hardware licensing agreements were awkward, shortsighted and restrictive -- Apple never quite seemed to commit. And the third-party developers, without strong Apple support and with high licensing costs, couldn't find a way to offer products significantly different from Apple's own. After Jobs' return, he decided to end the licensing experiment. His rationale: It had begun too late to really make a difference, and the clones were cutting into Apple's own sales instead of expanding the market. In late 1997, the whole shebang wound down, with Apple buying some of Power Computing's assets for $100 million.