Frequently, people who write serial stories--TV shows, series of novels, etc.--write themselves into a corner. The most notorious example is Dallas, which got into such a ridiculous snarl of plot twists that the producers decided, what the hell, most of what happened up to that point was all a dream. (Reboot.)
Can Superman defeat The Complicator?
Several years later, DC Comics, which owns the license to Superman, concluded that not even the Man of Steel's awesome powers could help him escape from the convolutions and complexities that piled up over decades of monthly publications. There was too much backstory for new readers to make sense of what was happening in any particular issue.
Not only was there a man of steel, but also a dog of steel (Krypto), a teen of steel (Supergirl), a chimp of steel (Beppo), a Superman from a parallel universe...And we're just talking about the good guys. Cataloguing the bad guys was an even more Herculean labor (perhaps as difficult as bending steel in your bare hands.) There were even several different types of kryptonite, each of which screwed with Superman in a distinct way.
This "continuity" also made it increasingly difficult for the writers to do anything new or different. Any story idea might be improbable or impossible, in a universe populated by a bazillion heroes and villains, plus an incalculable number of other complicated factors. ("Oh, right, if Superman did fight Eclipso, he'd lose, because Superman's vulnerable to magic. But why then wouldn't he just ask Dr. Fate to kick Eclipso's ass?")
Something had to give. DC's solution was to start over, wipe the slate clean, ignore everything that had occurred in the decades-long Superman storyline up to that point. The writers now had perfect freedom to do anything they want, as long as they stayed in the spirit of Superman. Lex Luthor was still a villain, but rather than being a crazed inventor, he became an amoral, power-mad industrialist.
Can Superman escape the Support Matrix?
What's the relevance to the technology industry? The common trait they share, particularly in software, is the growth of complexity to the point of unsustainability. It's not the inevitable outcome of ongoing development, but it can easily follow that path.
Every new version of a product, it gets more features. Add a knob here, a dial there, and the changes look pretty easy to digest. Add up these enhancements across several features, and you can create a product that's omnicapable, but nearly impossible to figure out on your own. Everyone likes to point to Microsoft Office as the best example of this point, but I can think of many others, too. (Some name-brand content management applications suffer from this malady, as do some multimedia editors I've used recently.)
Unfortunately, rebooting is not really an option. Not only do you risk losing customers if you cut too many features, but you're still contractually obligated to support people using the old product. (The phrase "legacy code" is as cringeworthy to development teams as "producer's actor nephew" is to movie directors.)
Still, it's important to regularly perform the mental exercise of cutting features you already have. Which didn't prove to be as important as we thought? What particular kind of customer cares about this feature, and why? It's easy to keep adding things; it's harder to remove them.
If you can't really cut back, at the very least, you can have an entry-level product. No, not some crippled application that doesn't solve anyone's problems, and is just a point of leverage for getting people to upgrade the real application. Instead, create a product that is useful, with a smaller range of features, to a particular audience that doesn't strictly need anything more. Go ahead and build more for other audiences, but always give someone the option to start with less. Let them decide for themselves if they want more.
The basic product becomes a lot like the DVD collections for some TV shows that focus on a particular plot element. Someone can watch the alien invasion episodes of The X-Files, or the Angel episodes of Buffy, and still enjoy that piece of the much bigger storyline. If a product team can't define the minimally necessary product, then they don't really understand their audience, and will face difficulties expanding it.
P.S. If you think that the fight against the cancerous growth of software features is harder than avoiding plot complexity in comic books, think again. The Superman comics have, in 20 years, again collected a lot of complex, interconnected, and unfamiliar story elements. Who the heck is Dominus? What's all the hoo-hah about this Doomsday guy? And how many times can Lex Luthor die, or at least appear to die, without the next death scene becoming unintentionally laughable? At least, in the technology industry, you don't really have to start all over again.