Why are our projects predictably unpredictable?
The 2012 Wilson Research Group Functional Verification Study shows that around 67% of our projects are late, continuing a trend going back at least ten years.
As an industry we are pretty bad at estimating how much time it will take to complete a design unless we have previously done something very similar. Yet many projects involve substantial changes: implementing new protocols, incorporating new versions of IPs, increasing the number of CPU cores, implementing new power management schemes, etc.
We know there are going to be surprises, but being able to spot problems early enough to take corrective action is difficult: many projects use design teams located around the world with different tools, methodology, and even design goals. Getting timely, accurate information is a challenging problem in itself.
This analysis is reinforced by a recent design management survey showing the top requirement from project management was for global visibility.