Building the Next Generation of Financial Infrastructure

Conway's Law:

>Organizations which design systems are constrained to produce designs which are copies of the communication structures of these organizations.

In my experience this is very true at banks. Perhaps it is truer at banks than anywhere else - although I suspect as much, I can't be sure. Bank software is basically about implementing in code, processes that were developed to be applied solely by humans 100 years ago or more. It's all the same stuff. There is nothing new. It's so bad that, after an organizational change, even if there is no logical reason that this change should affect the software running the bank at all, it will anyway, because some existing piece of technology was the old guy's pet project, and so the new guy needs his own pet project. And all projects are pet projects. And so on.

And that's just on the tech side. Of course, on the business side, technology is a cost to be grudgingly paid out and worked on, at best. And once the technical debt gets to be unworkable, someone comes up with the great idea to hand everything over (or in pieces, whatever) to a vendor. But then eventually the cost structure there gets too onerous as well, and so the business tries to replace it in-house again, after spending the last decade clearing out all their in-house expertise. And so the cycle continues.

Some banks are worse than others, but if this article has you thinking that maybe at Goldman Sachs, or Credit Suisse, or Morgan Stanley and Deutsche Bank, technology isn't an absolute shit-show, well think again. And these places run the world economy while their technological infrastructure is held together by duct tape.

The article is right, of course, that banks will still be around in 20 years. Obviously. But, the big banks have known they need to get their act to together for, at the very least, forty damn years. And not a single one of them (not one!) has done it. And they never will, because they are dominated by a culture going back two hundred years that will not ever let engineers anywhere near the levers of the business, which is where they need to be, and for that matter will most of the time not invest profits back into the business when they can funnel it into their own bonus instead.

If you want to disrupt banks, start a bank! Start a bank where you can do things and enter markets that the big guys can't, because their tech is half-assed and they aren't creative enough to do anything about it. Finance is in need of disruption, and the potential windfall to pulling it off is enormous, but you aren't going to get there by selling to existing players.