After working on a few different innovation teams at Amazon, I've found that innovation at an enterprise level can generally be classified as one of three categories - continuous innovation, discontinuous innovation, and big vision ideas. My mental model is the following:

1. Continuous Innovation. Incremental features within the existing business verticals. This is your atomic unit of innovation and tends to be the most common form across all business types and sizes.
2. Discontinuous Innovation. Most people think of this as disruptive ideas or new-to-world products. Through an enterprise scope, I see it simply as a new business vertical (new-to-us, not necessarily new-to-world).
3. Big Vision. Long-term product strategy for the existing business which takes about 3-5 years to achieve. This also requires 1-2 year continuous innovation strategy and support from various teams within an org/company.

I was fortunate to have worked in the retail org Concept Lab which evolved through different stages. We did everything from exploring new technologies and disruptive opportunities to smaller incremental features; from developing big vision ideas to crafting the near-term strategy. And with Amazon Glow, I was part of a recent new-to-world product launch, seeing the idea go from prototyping all the way to production. With each team, there were a lot things we did right but also plenty of learning opportunities; after some reflection, here are some thoughts on what I think are critical for success for each type of innovation. While the points aren't mutually exclusive and there are certainly edge cases where they don't apply, these are what I consider success criteria for the general case.

1. Continuous innovation requires cultural investment at the executive level.
Employees hired to run the existing business are closest to the problem/opportunity space and are best-positioned to identify improvements. A dedicated innovation team brought in to incrementally innovate is poor resource utilization at best and degredation to team morale at worse. Instead, it's better to create a culture that enables everyone in the org to innovate. The biggest innovation blockers are teams either heads down maintaining an existing system or implementing an existing roadmap, so much so that there isn't space to explore new ideas. Beyond having the space, time, and autonomy, team members need to develop the muscles to ideate and agility to validate concepts; to feel empowered to try and have a safety net to land on when things fail. And they need the infrastructure (tech capability, UX support, product approvals) to be robust enough so that they can not only think big but deliver big. This culture of innovation is often embedded within startups but is an absolute tectonic shift for larger companies where leadership/executive-level mandate is critical.

2. Discontinuous innovation requires proper funding and should be intentionally detached from existing business verticals.
That's not to say the catalyst for disruptive new ideas can't come from someone within the current team. However, the new initiative should have a clear separation from the existing business which requires proper funding. The problem is that the branches which support the current business may appear helpful at a glance but turn out to be thorny. Some examples include: finding a "home" prematurely, morphing the idea to fit within an existing business unit; enforcing current architectures (especially tech infrastructure/capability) as dependencies rather than treating as new opportunities with potentially a more optimal solution. It's better for these ideas to germinate and grow without the existing frame. One strategy is to control this at the seed level by bringing in outside expertise, cherry-picked to tackle a specific domain foreign to the company. This detached models helps retain focus and product differentiation up to launch, but post-launch introduces a unique set of problems with scaling and integrating with the company ecosystem which I won't get into here.

3. Big vision ideas and long term product strategy requires both healthy funding and a strong culture of innovation supported by a dedicated innovation team.
Again, these are roadmap strategies that take ~3-5 year to realize the full vision and requires every team within the org to execute on. This is where a dedicated innovation team can provide the most value - by tackling the big, non-incremental problems. There's definitely a danger with the innovation team being viewed as elitists which is counter-productive to both the mission and the culture of inclusivity. It's crucial to get partner teams bought into the 1-2 year strategy, and it starts by carrying them along the journey so that ownership is shared. For this to work, the innovation team's charter doesn't stop at the big ideas; they should also be responsible for facilitating the culture and enable teams within the org to continuously innovate along the roadmap.