Thursday, July 31, 2014

$840 million...

$840 million...  That's how much the federal government admits to having spent on the ObamaCare web site.  That is an absolutely incredible figure.  To anyone from the software industry (as I am), it is instantly clear that there is something terribly wrong here, because that number is at least an order of magnitude too high, and probably approaches two orders of magnitude too high.  If I were to guess at a “reasonable” figure for that web site's development, it would be something more like $10 to $20 million – and even that would be making allowances for government inefficiency.

Let's break that down a little bit to see why I immediately know that figure can't possibly represent a well-run project.  Let's assume we're hiring only American software engineers (which, it turns out, was not the case).  Let's assume we're hiring only well-qualified, experienced programmers (which, it turns out, was also not the case).  A guesstimate for the fully burdened cost (that is, salary plus benefits, etc.) of each such engineer would be about $160k annually.  Now let's assume one manager for every 5 engineers (there were actually nearly as many managers as engineers), and that their burdened cost was $200k annually.  There were three years of work.  How big a team would $840 million buy us?  That would be 4,200 software engineers, and a total of 12,600 engineer-years of work.

How much is that?  Well, it's more work that it took to build Windows, Microsoft Office, and OS X combined.  Each of those projects is clearly larger and more complex – and vastly more technically challenging – than the ObamaCare website and it's back-end. 

From another perspective, though, it's not too surprising.  Large software projects are notoriously difficult to manage.  Like any veteran of the software industry, I can cite numerous examples of failed projects from my own personal experience.  Sometimes the failures are technical, sometimes business, sometimes management – or any combination of the three.  In a business, however, failures tend to get recognized and fixed relatively early, before the cost (relative to benefit) gets out of hand.  When the government is running the show, none of the usual business incentives to find and fix a problem early are in place.  Instead, there is a perverse opposite incentive, to keep the bureaucracy running – so the government's classic behavior in the face of a failed project is to throw more and more money at it.  I'd be willing to bet you that that's what happened here.

Very sad, the way my tax dollars were totally wasted here.

Virtually every development executive – no matter how bad – in every software company I've ever worked with would likely have done a better job, if given business-like incentives.

Pardon me for a while, please.  I need to go cry...

No comments:

Post a Comment