Intuit's founder, Scott Cook, believes that success can actually be dangerous to the company. At a seminar with Harvard Business School faculty, he said that "Success is a powerful thing, it tends to make companies stupid, and they become less and less innovative."
According to Harvard Business School Working Knowledge, Cook argues that companies need to adopt the lean startup model pioneered by Eric Ries. That means "launching as quickly as possible with a "minimum viable product," a bare-bones creation that includes just enough features to allow for useful feedback from early adopters. The company then releases a quick succession of product upgrades, forming hypotheses and conducting experiments with each new version along the way."
Getting feedback early and often means the product improves quickly, picking up new users along the way.
That's the polar opposite of the way many successful companies come up with new products. They spend years developing things, many of which are only minor improvements on what already exists. Once something's set into motion, it has a great deal of inertia and can take a long time to stop, even if it's not working.
That's an easy way for a company to stagnate. However, adopting something like the lean startup model takes a significant cultural shift. It's not easy to take product teams used to taking years and get them to take ideas from birth to execution in months. It also takes creating a culture that's OK with failure and starting again, which is the opposite of what you see in many large companies.
Some companies choose to have a small part of their company, like a skunk works, to constantly try to innovate. But when it's a small part of the company, you see small benefits. Far more effective is forcing a whole company to constantly experiment and innovate.