Because of the natural human tendency to hide problems information systems often fail before they are even half way to completion. For this same reason it is difficult to trust any statistics that are given out !
Some posts i have read suggest that we should look at the building of information systems the same way we do the building of a twenty story office. Each stage of this must be meticulously checked and rechecked to avoid disaster.
If we were to keep the building of said offices in mind when we were to build our information systems then a lot of our problems would go away. The illusion that there is a problem with the information system if we are not writing code in three months is one we need to leave behind and strive for a more open and sensible approach when constructing a system.