In my career, I have seen a common pattern where systems are grown around limitations on the system architecture. A good example - we don't collect or intelligently use every click from a user because there isn't just enough disk space. With disk space becoming so cheap, why don't leading internet companies figure out a way to collect all that information? By itself, it may not be valuable but in context of products for sale and different product placement and marketing techniques, capturing behaviour can give you the edge over your competitors.

I propose and imagination is the limit way of development where analysts keep up to date documents on software and business model improvements based on incremental improvements to the architecture. Here are my pains. What could double the processing power do for me? What are historical systems conversions that I'd have to do. Could my historical systems/data be used better with more system resources? What are the opportunities for the go forward processes?

You may have to drop your user behavior trails in the short run but there is nothing stopping you from capturing your business model and software and data capture ideas. Save it for a time when you can do something about it. When that time comes, execute, execute, execute.

Here's the system at phase I. We are fully utilizing our resources to make money for the company. Now... if we had more hard drive space... if we had more cpu.... if we had more programmers.... we could do these 100 extra things... here's how it would work better than it's working now....