Five years and 500 million dollars. Thats what it took the California Public Employees Retirement System to finally launch its new computer system my|CalPERS.
Please let that number sink in for a moment.
Five. Hundred. Million. Dollars. And thats the green kind of dollars made in the US of A.
In what planet is $500 million a reasonable amount of money for a pension fund to pay for a one-off computer system? Seriously, lets take a second and think about how that $500 million could be deployed in other ways...
Imagine you took that $500 million or lets start with just half of it and used it to launch a NEW company in Silicon Valley with world-class technologists and an objective to build data management and benefits administration software for pension funds. The $250 million dollars youre going to put into this new company could buy you 400 engineers of development time for five years. You could build a lot of cool software with that amount of people. And if you can convince other pension funds of the value of this venture, you could share the costs.
And you wouldnt put ALL the money in at once. You would set milestones and have tranches and drawdowns based on actual accomplishments...you know, like VCs might do. And the money would go almost exclusively towards the best engineers; no Aeron chairs allowed (though those big bouncy ball chairs are OK).
I think such a company could capture the employment allegiance of talented big data engineers from the likes of Google and front end GUI engineers from companies like Apple. How? By giving upside in the form of equity and options in a company that has $250 million in secured financing and a cornerstone client by the name of CalPERS (which would provide another $250 million in secured revenue for the first ten years). So recruiting wouldnt be challenging. There are plenty of Stanford engineers that would jump at the chance to work for such a company; to try to solve some of the challenges of ''big data'' in novel ways.
Are you following me here? The path that Im describing is one that would see CalPERS use the money it spent on a software package to instead build a business that offers the same capabilities to other pension funds and thereby create a feature set that no one of the pension funds could have afforded to develop on their own. It might be risky, but so long as the lead pension fund client gets the data management facilities it needs at the end, does it matter?
And if CalPERS did manage to build a world-class company, it might help other pension funds from having to throw hundreds of millions of their own money into this black hole. The new company could sell my|PENSION to funds around the world, and CalPERS would then have a majority stake in a profitable growing business instead of a capitalized cost. If its going to be on the balance sheet anyway, why not have it as an investment?
You think Im crazy. Youre not alone. But, this time, Im certain this is doable. For example, go and look at this company, which offers a high-powered data engine to governments (e.g., the FBI) and finance companies (e.g., JP Morgan). This company is a useful comp for the type of company Im thinking of in that it mixes huge data sets with government bureaucracies and finance. Since the companys inception in 2004, it has raised roughly $300 million in funding (which, as a reminder, is $200 million LESS than CalPERS spent on its single data platform). And today the company is valued at between $2.5 and $3 billion.
In short, I see a pension fund spend half a billion dollars on software, and I start to lose it. I wonder how that money could have been put to better use. And I know I know that CalPERS could have solved its own problems and those of countless other pension funds with a little bit of ingenuity and a willingness to think outside the box. It could have taken that $500 million and turned it into a $5 billion data and benefits administration company. Why didnt it?