The cases of IBM and more arguably Microsoft and Oracle tell us that if you reach a certain scale, you can continue to coast for a very, very long time after you lose relevance. Microsoft might be an example of having the glide time to regain it depending on how Windows 10 and Azure go over years, but it's far easier for a large corporation to spend billions defending their shrinking turf than to make the big changes usually required to regain it. The problem is that defending your shrinking turf can show positive revenue numbers for quite some time, so it all looks great while it happens...