Diary

The myth of 100% systems availability

Aside from the fact that the numbers are interesting to consider, it struck me reading this piece in the IBD this morning how similarly my thought process works from one context to the next. It’s reported here that the number of those without health insurance rose from 46.3M in ’08 to 50.7M in ’09. My first thought was, compared to what baseline? Sure enough, that’s where the author went.

In my business life I have been told by clients, literally hundreds of times, that they “must have” 100% I/T systems availability. In practice no one does of course– things break and there’s little, realistically, you can do about it. But you can get pretty close. I always ask them first what they’re getting now and somehow even after years of having this discussion, I’m somehow always surprised at the answers. Most common answer? “I don’t know.” Most of the numerical responses… which is to say probably all of them … are also ultimately amended to “I don’t actually know.”

It hit me in considering public policy issues I’m also asking myself, against what baseline? In pretty much every case, the baseline in my mind is what the Constitution allows. It doesn’t allow for the gargolye we lovingly call Obamacare of course. Then you consider the definition of the desired end state and once you arrive at that, you can assess what it might cost and whether you can even get there from the baseline. My current client would say they have to have 100% availability because it costs them X millions of dollars an hour in lost business to be down. Even assuming X is an accurate number (whatever that number is and whomever the client is, it’s usually wrong), in this case when you tell them chances are they can actually tolerate 92% because they probably sell close to $0 worth of insurance between 1 and 3AM any given morning, their point of view changes.

Clearly in the case of Obamacare, you can’t get there from the baseline, but that hasn’t stopped ‘progressives’ from trying because they haven’t examined the real meaning of the desired end-state and they haven’t considered the actual practical ramifications of doing so. So that 50.7M represents 16.7% of the population. That appears to be bad — a prolonged system outage of some sort usually provokes the availability discussion– but is it really when compared to the total trend line? As the IBD points out “[o]bserved on a chart, the rate is a rather straight line that bumps up a bit in 2009 as the economy went cold.”

Looks bad anyway though, right? Well, back to the baseline–

While the ratio of uninsured remains steady, the number of Americans with coverage has increased by 22 million over the last decade. At the same time the number without coverage “only increased about 5.3 million.”

For this, the country has been saddled with socialized medicine that will cost trillions and erode the quality of treatment?

This is exactly the sort of thing that causes clients climb down out of the 100% availability tree. When they discover the massive I/T budget increase they’re going to get, the disruption implementation will cause, realize they won’t get stone-cold 100% availability anyway, that they can actually tolerate some down time because they don’t understand what down time really means, they look for a more practical answer– one that fits with their amended desired end-state. The insanity of Obamacare isn’t much different than a client that refuses all application of common sense to their problem and insists on “100% availability.” My whole life’s experience tells me that even given an infinite supply of time and money, ain’t gonna happen.