Whatever happened to the development cycle?

I've been thinking about the sequence of events in large retail oganisations when a decision is made to implement an in-house point-of-sale system. Almost immediately, project managers, business analysts, end-users and developers flurry around frantically to follow something that resembles a software development life cycle. Heated discussions take place to decide the system scope. Sales staff are inundated with questions about their system needs. Detailed specifications are delivered to developers, and the system is tried and tested before rolling out to the users.

Those of us in the real world, however, know how often we stray from the ideal. How many times have systems been rejected by users because they were not included in the process? Because the system didn't meet their needs? Or purely because their current manual processes were actually good enough to do their jobs?

Now consider a decision to include the Internet as a sales or marketing tool. Why is it any different to implementing an in-house point-of-sale system? What is the Internet if not a point-of-sale system with a different end-user? When you consider that in the online environment, the new end-user is an unknown entity using the system in an unknown context, are the chances of rejection not that much higher? Perhaps that's a clue to why online conversion rates of 1% have become acceptable standards.

I'm not saying that every web initiative deserves the full consideration that goes into normal system development. All I'm saying is that perhaps it needs more consideration than it's given most of the time.