Over at BankerVision, James Gardner, heads toward blasphemous territory when he suggests that in some cases, paper based manual processing may make more sense and cost less than grappling with IT complexity.

IT systems are now so expensive we are at an inflection point. The point where doing things manually is more economically efficient for certain higher value, lower volume products than doing them automatically.

Paper as an innovative solution requires political fortitude on the part of the proposer. It is almost always unpopular. We've all been so indoctrinated in thinking that people based processing is bad we no longer even put it up as an option any more.

But look at its advantages. It is quick to get going. Quicker than practically any kind of IT. It is robust, and can recover from unexpected failures with ease. And it is cost effective for startup projects that don't have all that much volume and don't know exactly how things are going to work out at the beginning.

IT has none of those attributes. And I doubt it will get them any time soon.

Consider this: you can put paper behind a web site or any other self service channel and have the most fantastic experience you can imagine.  It is therefore a wonder to me why web startups invest so much in their backoffice IT when they could probably get by with a beautiful user interface and a bit of elbow grease.

Elbow grease indeed! I can't endorse paper, but I am all in favor of piloting new processes using Excel and Access behind the scenes to fine-tune requirements and test business cases hypotheses.