Ben Schaffer's mid December 2006 article Open Standards for Online Advocacy Tools focuses on software used in fund raising and politics, but is about inter-operability issues equally applicable to business domains such as health care or pharmaceutical sales.
Here's a critical bit:
Because the choice of software is so hard to change, there is tremendous pressure during the initial sales phase for a vendor to convince you of the superiority of its software, but reduced incentive for its software to continue being the best over the rest of your relationship. (Cell phone plans are a good example of this same phenomenon.)
This also puts the burden on the client to choose software on Day 1 that will be appropriate for them for every stage of their unknown future growth, and that will anticipate all of their unknown future needs. Ask any commodities trader; it's not easy to do.
For this reason, many organizations are today stuck using software they have already outgrown.
Alternately, you might pick a specialised software package that is geared to one particular aspect of your campaign. For most organizations, fundraising is understandably central, so the fundraisers get to choose the software used. If this software works for other departments, great, but if it doesn't they are precluded from choosing their own tools -- unless, once again, they simply keep parallel data.
The ramifications of this approach are similar: you're locked into your choice unless you abandon it totally or run multiple apps with no communication between them.
I think he missed the role customisation plays in locking software licensees into their own pasts, but overall, that's as good a statement of the generic stove piping problem as I've seen. His description of open, standards based, inter-operability is even better:
To continue the example of cell phone plans, many cell phone service providers were against number portability because they might lose customers. Other providers looked at it as an opportunity to attract new customers.
Data portability and sharing between applications is the same situation. I believe there is more opportunity for growth because of it than there is likelihood of failure. Far from being a chance to lose customers, it should be seen as a chance to create converts. Because ours is a young and innovative industry, I believe we can compete with features, price, and all the usual methods, without holding clients and their data hostage.
With increased interoperability, we enable new possibilities. A small team of smart people can could quickly develop a unique tool that solves a particular problem, which could be used by clients in tandem with their existing software. In this scenario, the client obviously benefits from the tool, but both the new and existing software providers benefit from the fact that the client does not have to choose between them. New developers do not have to take clients away from established developers in order to gain clients for themselves. To my way of thinking, this is an even bigger advantage for existing developers than it is to new ones. To some extent, it precludes an arms race and allows vendors to focus on their strengths.
He works in the campaign software business, but the issues are every bit as real, and the solutions every bit as applicable, whether your concerns involve ERP/SCM applications for Fortune 100 companies or a child's choice of word processor to meet school submission requirements.