Become the thing that replaces you
I asked Little Miss MySpace what happens when something new comes along... when someone else makes a MySpace-Killer. Skyler said, "Why does it have to be someone else? MySpace can just become that (whatever 'that' is)." She knows nothing of the business or politics of MySpace--she's simply a passionate user. And she's never read The Innovator's Dilemma. But she still has a point: why shouldn't we be the ones to build our own "killer"?
Whether we're trying to innovate around our existing products and services or trying to find a completely new idea, we have to back up to the meta-level rather than focus on implementation. Obviously implementation matters... a lot. But implementation of what? Why build a better XYZ if all that matters to users is the Z? What if XYZ is just one way to give users what they really want--JKL--and there's actually a much better way to help them do JKL? A way that makes XYZ unnecessary?
If we're not careful, we can take our existing success and misattribute it to an implementation detail that was never important.
Or worse... we can misattribute our success to something that's actually a problem but that users managed to cope with. Right up to the time we upgraded that thing-they-never-liked to give it a bigger role.
Yes, this is just another one more of those DUH topics, but as with so many others--it's too easy to get sidetracked by either our own success or the success of someone else's product or service that we're trying to build a better version of. And that's one reason why trying to reverse-engineer the success of a product is tricky. We get stuck rationalizing why some implementation detail is important, when it may be nothing more than noise.
An example of outstanding implementation that ignored the meta-level
The now-dead Purple Moon software company was the result of millions of dollars and years of research at Paul Allen's Interval Research think tank. They had finally found the secret sauce to getting girls--the great untapped market--into gaming. By the time the company's first product launched (1997), they knew just about everything you could know about what young (10-14 years) girls wanted and how they differed from boys. So they took their exhaustive and expensive research effort and created the ultimate implementation.
The implementation was awesome... beautiful graphics, clever characters and story, slick marketing, and a world-class leader who many of us still practically worship, Brenda Laurel. If anythng could finally bring girls to the games, it would be the perfectly-pedigreed Purple Moon's first game, Rockett's New School.
Except it sucked.
At the meta-level.
Because what is the meta-level for a game? Oh yes, fun. Purple Moon got the individual implementation details right, and applied all that they'd learned from the research, but forgot the forest. Fun.
Skyler was an early beta-tester, and had been looking forward to the game I'd been hyping for so long. But the first thing she said when she read the overview document was, "Why would I want to play a game about a girl trying to fit into a new school? HELLO! I've DONE that in real life and it wasn't fun then. I'd rather play Blobbo."
[Warning: gratuitous kid photo... this is Skyler today]
[Fortunately, she reads this blog only twice a year--she'd kill me if she knew I posted this. Let's just keep it our little secret.]
Granted, Skyler wasn't the typical pre-teen. She didn't do Barbie. (She would have given a kidney for her My Little Pony collection, however). But still, when you strip away Purple Moon's research and implementation details, Occam's Razor applies: Just. Wasn't. Fun.
With our books, that meta-thing is learning. And if we get off track by focusing on and EQ'ing our implementation details without remembering that, we're sunk. So if we try to figure out our own "killer", we'll do it only by staying true to the meta-level forest.
Many of us are creating products or services where the barrier to entry for a competitor is not all that high. The only thing we have to really protect us is a willingness to throw out even our most successful products in order to build a better reflection of what matters to users at the meta-level. And that might look nothing like our current, successful product. Keeping focused on meta-levels is also the key to avoiding being trapped by fads or fashion. Fads and fashion ("rounded", "glossy", "extreme", "twittery" [sorry, couldn't resist ; ) ] tend to be implementation details, not meta-level concepts ("have fun", "kick ass", "be smarter", "have more time in flow", etc.)
Finding the meta-level
The best trick we know for finding the meta-level is to play the five-why's / why-who-cares-so-what game. Ask your users (or even just yourself) what's important about a product. When they answer, ask, "Why?" When they answer that, say, "So?" and when they answer that, say, "Who cares?" and keep going until you get to the heart of it. (And if you haven't played this before, most people stop WAY too early and miss what matters the most.) Only then do you discover that this feature the users--and you--believed to be meaningful was simply a tolerable way to do what they really wanted. When they say that X is important, dig deep enough and you might find that it was only because X let them do Z, and that there's a much better way to make that happen.
Again, I know we all know this. But it's so hard to do, and the more successful your product or service, the harder it becomes. "Don't mess with success" is often the biggest barrier to becoming your own "killer".
A prominent tech book author wrote on a public forum, "Your Head First books will be fine just until the next hot new thing comes along to replace it." I said, "Yes, and that's why I want to be the one to replace it."
A big hat tip to Lassi Etelaetalo of Finland, who won the 
On 
As part of a series of "Innovation Wednesday" blog posts about innovation, Yvonne DiVita of the 
