Buying any kind of software requires a great deal of thought into why and what it is going to be used for. Intelligence software is no different in that respect. Please share best practices in this area and what the key points are to think abut before deciding what to implement. What must be included and what is not so important? Should everything come from one vendor or should you build it on your own?
The report referenced here is one of the most thorough CI software case studies ever written. Most of the 20+ pages are direct quotes from a CIO at a European Pharmaceutical firm. He talks about the research and selection process, as well as the planning, deployment and training process.
Indeed, procuring software requires a good understanding of the following concepts:
- Why do I, my team, my company need a software for intelligence activities?
- What does it cost me, my team, and my company not having a software today?
- Which value should this software bring now and in the future to me, my team, and how will it positively impact information flows and processes in my company?
There are some good answers from other posters here. The last question you raise about building your own solution instead of going for something provided by a vendor has it's own interesting aspects. We did a white paper on this a couple of years ago: http://www.comintell.com/website/downloads/White%20Papers/S0130-F%2...
While some people are aware that it might be costly to do their own development, fewer realise that home cooked solutions tend to underperfom as well. We have some references to that in the white paper. So, in answer to your last question, I would prefer a vendor solution or a combination of vendor solutions to building my own.
I think that there an interesting White-Paper from Digimind attaching which are the main features for a CI software. This is a good starting point for a check list for doing a benchmark and comparative cost-features.