Very early at the start of the selection you should determine what the exact criteria are for selecting a certain piece of software. You will ask the vendors questions. They will answer you. And then what? Look who has the nicest handwriting? You have to think about this up front.

It's even worse… the questions you will ask are a result of the criteria you want to judge. How else could you say something about your software selection criteria? Ok, let's have a look at some criteria I personally use. There is no real definite list of selection criteria. They can very from company to company, from individual to individual. I included the following selection criteria for illustration purposes (and remember, I'm a prototype "early majority").

Stability of the vendor

I work for a large corporation. If we buy a piece of software it will probably be part of our primary operation for more then 4 years. It would be a big problem if a supplier went belly-up. We would have no more support, no more security, our life-line would be compromised. So, I would be looking for a vendor with some great stability.

Stability can be found in companies that operate already for a long time; that have a large customer base; have a steady income and revenue. Stability is also determined by company size (more people, means better equipped to handle people leaving), and, by looking at the people in charge: are they in charge, how are stocks divided, etc.

Proven track record within our industry

It's a fine line between amateurs and professionals, if you can judge people only superficially in very short time. You don't want to have amateurs fiddling with your company. Look for a proven track record within your industry.

Fit-for-purpose (least possible changes to the standard software as possible)

Changes to software are always a concern. They are difficult to specify, to develop and in general to plan. To reduce the amount of specific changes made for your company, is a way to reduce a project risk. You have a lower risk of schedule and budget overrun. If that's very important, you have to look for the software that comes close enough to what you actually want: fit-for-purpose.

Mature technology

I don't care how a certain piece of software is constructed. I just want to be sure that I'm not a kind of test case, a field study to see how "this cool stuff actually works". Having a new system is difficult as it is… using some technology that doesn't work quite yet, doesn't make it easier.