Old for New
The sheer volume of Access databases (actually, I came across five critical line of business apps running on FoxPro ver 2.6b, non-networked, users share data via floppy disks!! For you youngsters, they're like CD's or DVD's but smaller and less shiney!) and cheap , unreliable custom apps, developed by Chuck in Dallas, prevents anyone from thinking about a refresh strategy simply because no-one wants to have a 5 year long headache (and that's just from considering the data migration strategy).
When we look at the lack of adoption of TRUE Enterprise Architecture (i.e. a representation of a business combined with a realisation of technology produced by the business and owned by the business, NOT a bunch of documents and processes written by Mister Anonymous and owned by Mister Nobody) it is easy to see why, with so many wrongs to right within the technology estate of a large organisation, EA is either overlooked ("we have a 38 year old mainframe application which runs our core line of business functions! We need to replace that, not draw pretty pictures of what we all know!") or seen as "nice to have" ("Yes Archie, we can see you have a point but the Head of IT and the CEO are meeting to discuss outsourcing all support functions and I'm not sure they will want to see your pretty pictures!"). Cynical I know, but this is the crux of the issue. EA is owned by the business! Managed carefully and invested properly, it quickly becomes the realisation of a business strategy. All the technology headaches are eased by having a clear business model / case from which to build systems and solutions.
There is another twist to this folly that has emerged; the proliferation (ooo, twice in one posting) of consultancies who have risen to explain the complex phenomenon that is EA! We now have so many methodologies that it is almost like starting over! You would think that these consultancies have smelt the cash to be made out of confusion and complexity.
Without the slightest hint of cynicism, I say: