It's not just the one person but also the process and tools used. Intel's CPU design process tends to be stuck further back in the past than AMD's or Apple's. Instead of taking advantage of design automation Intel tends to throw more bodies at the problem. That doesn't scale well. Keller was brought in because of what he did at AMD and Apple to streamline the process and use more design automation. But Intel tends to hire people right out of school and they tend to stay around for many years so they've never been exposed to how things are in other companies. I have a suspicion that Intel has most of their US CPU design activities in Oregon not just because it's a cheaper location than SV but also because it's a lot harder for employees to switch companies - they'd have to move to the Bay Area or Austin and that's a tough move for people who have lived in Oregon for many years.
So much of what governs the success of large engineering projects is the processes that govern decision making: from the actual design of the thing you're trying to build, to staffing and resource allocation, norms and processes that govern how work groups interact and collaborate, etc. Adjusting those is the most important lever someone like a VP has but it takes time. You can make changes quickly but things will tend to break and sometimes the thing that breaks is the leader pushing the change.
